Volume 42, Number 2, APRIL-JUNE 2020 IEEE Annals Of The History Of Computing

Citation preview

Volume 42  Number 2  

Image Courtesy of Eileen Buckholtz

www.computer.org/annals

APRIL-JUNE 2020

Queens of Code

IEEE Computer Society Has You Covered! WORLD-CLASS CONFERENCES — Stay ahead of the curve by attending one of our 200+ globally recognized conferences. DIGITAL LIBRARY — Easily access over 780k articles covering world-class peer-reviewed content in the IEEE Computer Society Digital Library. CALLS FOR PAPERS — Discover opportunities to write and present your ground-breaking accomplishments. EDUCATION — Strengthen your resume with the IEEE Computer Society Course Catalog and its range of offerings. ADVANCE YOUR CAREER — Search the new positions posted in the IEEE Computer Society Jobs Board. NETWORK — Make connections that count by participating in local Region, Section, and Chapter activities.

Explore all of the member benefits at www.computer.org today!

April-June 2020 Volume 42 Number 2

5F

rom the Editor’s Desk

Gerardo Con Diaz

Published by the IEEE Computer Society

6 P

isa, 1954–1961: Assessing Key Stages of a Seminal Italian Project Giovanni A. Cignoni and Fabio Gadducci

20 F

ake But True: Model Maker Roberto Guatelli, Science Museums and Replicated Artifacts of Computing History Silvio Hénin and Simona Casonato

33 T

he First Computer in New Zealand Brian E. Carpenter

42 O

nce FITS, Always FITS? Astronomical Infrastructure in Transition Michael Scroggins and Bernadette M. Boscoe

Departments Anecdotes ueens of Code

55 Q

Eileen Buckholtz

63 L

earning From Prototypes

Zbigniew Stachniak Interview n Interview with MICHAEL R. WILLIAMS David Walden

72 A

Book Review eryl Alper, Giving Voice: Mobile Communication, Disability, and Inequality. Cambridge, MA, USA: MIT Press, 2017 Nabeel Siddiqui

87 M

ISSN: 1058-6180

Image Courtesy of Eileen Buckholtz

Published by the IEEE Computer Society | www.computer.org/annals EDITOR IN CHIEF Gerardo Con Diaz, University of California, Davis [email protected] ASSOCIATE EDITOR IN CHIEF David Hemmendinger, Union College [email protected] ASSOCIATE EDITORS Mar Hicks, Illinois Institute of Technology Jeffrey R. Yost, Charles Babbage Institute DEPARTMENT EDITORS Anecdotes: David Walden ([email protected]) Biographies: [email protected] Events and Sightings: [email protected] Interviews: Dag Spicer, Computer History Museum ([email protected]) Reviews: Gerardo Con Diaz, University of California, Davis ([email protected]) Think Piece: [email protected] FORMER EDITORS IN CHIEF Bernard A. Galler, 1979–1987 J.A.N. Lee, 1987–1995 Michael R. Williams, 1996–2000 Tim Bergin, 2000–2003 David A. Grier, 2004–2007 Jeffrey R. Yost, 2008–2011 Lars Heide, 2012–2014 Nathan Ensmenger, 2015–2018 IEEE ANNALS OF THE HISTORY OF COMPUTING STAFF Peer Review Administrator: [email protected] Publications Portfolio Manager: Carrie Clark Publisher: Robin Baldwin Executive Director: Melissa Russell Senior Advertising Coordinator: Debbie Sims IEEE Computer Society Executive Director: Melissa Russell IEEE PUBLISHING OPERATIONS Senior Director, Publishing Operations: Dawn Melley Director, Editorial Services: Kevin Lisankie

Director, Production Services: Peter M. Tuohy Manager, Journals Production: Katie Sullivan Journals Production Manager: AndreAnna McLean, [email protected] Associate Director, Information Conversion and Editorial Support: Neelam Khinvasara Senior Art Director, Magazines: Janet Dudar COMPUTER SOCIETY OFFICE 10662 Los Vaqueros Circle, Los Alamitos, CA 90720 USA; +1 (714) 821-8380

EDITORIAL BOARD Gerard Alberts, University of Amsterdam Martin Campbell-Kelly, University of Warwick James W. Cortada, Charles Babbage Institute Peter Denning, Naval Postgraduate School David Alan Grier, George Washington University Thomas Haigh, University of Wisconsin Ulf Hashagen, Munich Center for the History of Science and Technology Chigusa Kita, Kyoto University Jennifer Light, Massachusetts Institute of Technology Elizabeth Petrick, Rice University Mark Priestley, Independent Researcher Brian Randell, Newcastle University Sarah Roberts, University of California, Los Angeles Corinna Schlombs, Rochester Institute of Technology Lee Vinsel, Virginia Tech CS MAGAZINE OPERATIONS COMMITTEE Sumi Helal (Chair), Irena Bojanova, Jim X. Chen, Shu-Ching Chen, Gerardo Con Diaz, David Alan Grier, Lizy K. John, Marc Langheinrich, Torsten Möller, David Nicol, Ipek Ozkaya, George Pallis, VS Subrahmanian CS PUBLICATIONS BOARD Fabrizio Lombardi (VP for Publications), Alfredo Benso, Cristiana Bolchini, Javier Bruguera, Carl K. Chang, Fred Douglis, Sumi Helal, Shi-Min Hu, Sy-Yen Kuo, Ming C. Lin, Daniel Zeng

IEEE Annals of the History of Computing (ISSN 1058-6180) is published quarterly by the IEEE Computer Society, 10662 Los Vaqueros Circle, Los Alamitos, CA 90720-1314, USA; phone +1 714 821-8380. IEEE Computer Society Headquarters, 1730 Massachusetts Ave. NW, Washington, D.C. 20036-4928, USA; IEEE Headquarters, 345 East 47th St., New York, NY 10017, USA. Subscribe to IEEE Annals by visiting www.computer.org/annals. Postmaster: Send undelivered copies to IEEE Annals of the History of Computing, IEEE, Membership Processing Dept., 445 Hoes Lane, Piscataway, NJ 08855, USA. Periodicals postage paid at New York, NY, and at additional mailing offices. Canadian GST Registration No. 125634188. Canada Post Publications Mail Agreement Number 0487880. Return undeliverable Canadian addresses to P.O. Box 122, Niagara Falls, ON L2E 658. Printed in the United States. Reuse Rights and Reprint Permissions: Educational or personal use of this material is permitted without fee, provided such use 1) is not made for profit, 2) includes this notice and a full citation to the original work on the first page of the copy, and 3) does not imply IEEE endorsement of any third-party products or services. Authors and their companies are permitted to post the accepted version of IEEE-copyrighted material on their own Web servers without permission, provided that the IEEE copyright notice and a full citation to the original work appear on the first screen of the posted copy. An accepted manuscript is a version that has been revised by the author to incorporate review suggestions, but not the published version with copyediting, proofreading, and formatting added by IEEE. For more information, please go to www.ieee.org/publications_standards/publications/rights/paperversionpolicy.html. Permission to reprint/republish this material for commercial, advertising, or promotional purposes or for creating new collective works for resale or redistribution must be obtained from IEEE by writing to the IEEE Intellectual Property Rights Office, 445 Hoes Lane, Piscataway, NJ 08854-4141 or [email protected]. ©2020 IEEE. All rights reserved. Abstracting and Library Use: Abstracting is permitted with credit to the source. Libraries are permitted to photocopy for private use of patrons, provided the per-copy fee indicated in the code at the bottom of the first page is paid through the Copyright Clearance Center, 222 Rosewood Drive, Danvers, MA 01923. Editorial: Unless otherwise stated, bylined articles, as well as product and service descriptions, reflect the author’s or firm’s opinion. Inclusion in IEEE Annals of the History of Computing does not necessarily constitute endorsement by the IEEE or the IEEE Computer Society. All submissions are subject to editing for style, clarity, and length. For submissions, go to www.computer.org/annals/write-for-us. IEEE prohibits discrimination, harassment, and bullying. For more information, visit www.ieee.org/about/corporate/governance/p9-26.html.

April-June 2020

3

PURPOSE: The IEEE Computer Society is the world’s largest association of computing professionals and is the leading provider of technical information in the field. MEMBERSHIP: Members receive the monthly magazine Computer, discounts, and opportunities to serve (all activities are led by volunteer members). Membership is open to all IEEE members, affiliate society members, and others interested in the computer field. COMPUTER SOCIETY WEBSITE: www.computer.org OMBUDSMAN: Direct unresolved complaints to [email protected]. CHAPTERS: Regular and student chapters worldwide provide the opportunity to interact with colleagues, hear technical experts, and serve the local professional community. AVAILABLE INFORMATION: To check membership status, report an address change, or obtain more information on any of the following, email Customer Service at [email protected] or call +1 714 821 8380 (international) or our toll-free number, +1 800 272 6657 (US): • • • • • • • • •

Membership applications Publications catalog Draft standards and order forms Technical committee list Technical committee application Chapter start-up procedures Student scholarship information Volunteer leaders/staff directory IEEE senior member grade application (requires 10 years practice and significant performance in five of those 10)

PUBLICATIONS AND ACTIVITIES Computer: The flagship publication of the IEEE Computer Society, Computer publishes peer-reviewed technical content that covers all aspects of computer science, computer engineering, technology, and applications. Periodicals: The society publishes 12 magazines‫ژ‬ƌȄư‫ژז׏ژ‬DZȏɓȵȄƌǹȽ. Refer to membership application or request information as noted above. Conference Proceedings & Books: Conference Publishing Services publishes more than 275 titles every year. Standards Working Groups: More than 150 groups produce IEEE standards used throughout the world. Technical Committees: TCs provide professional interaction in more than 30 technical areas and directly influence computer engineering conferences and publications. Conferences/Education: The society holds about 200 conferences each year and sponsors many educational activities, including computing science accreditation. Certifications: The society offers three software developer credentials. For more information, visit www.computer.org/certification.

BOARD OF GOVERNORS MEETING ‫ژגא‬٫‫ דאژ‬°ƷȲɋƷȂƨƷȵ‫ژ׎א׎אژ‬ǠȄ‫ژ‬uƩkƷƌȄً‫ژ‬ÝǠȵǒǠȄǠƌً‫ژ‬Å°


revised ‫ڳת‬uƌɲ‫ש׫ש׫ڳ‬

EXECUTIVE COMMITTEE President: kƷǠǹƌ‫ژ‬%Ʒ‫ژ‬FǹȏȵǠƌȄǠ President-Elect: FȏȵȵƷȽɋ‫ژ‬°ǚɓǹǹ Past President: ƷƩǠǹǠƌ‫ژ‬uƷɋȵƌ First VP: ¨ǠƩƩƌȵưȏ‫ژ‬uƌȵǠƌȄǠ; Second VP: °ɲ‫ٯ‬äƷȄ‫ژ‬hɓȏ ‫ژژژژژ‬ Secretary: %ǠȂǠɋȵǠȏȽ‫ژ‬°ƷȵȲƌȄȏȽ; Treasurer: %ƌɫǠư‫ژ‬kȏȂƷɋ VP, MemberȽǚǠȲ & Geographic Activities: Yervant Zorian VP, Professional & Educational Activities: °ɲ‫ٮ‬äƷȄ‫ژ‬hɓȏ ‫ژژژژژژژژژژژ‬ VP, Publications: Fabrizio Lombardi VP, Standards Activities: Riccardo Mariani VP, Technical & Conference Activities: William D. Gropp‫ژ‬ 2019–2020 IEEE Division VIII Director: Elizabeth L. Burd‫ژ‬ ‫ژ׏א׎אٮ׎א׎א‬U---‫ژ‬%ǠɫǠȽǠȏȄ‫ژ‬Ý‫ژ‬%ǠȵƷƩɋȏȵ‫¾ژي‬ǚȏȂƌȽ‫ژ‬uِ‫ژ‬ȏȄɋƷ‫ژژژژژژژژژژ‬ ‫ژ׎א׎א‬U---‫ژ‬%ǠɫǠȽǠȏȄ‫ژ‬ÝUUU‫ژ‬%ǠȵƷƩɋȏȵ‫ٮ‬-ǹƷƩɋ‫ژي‬ǚȵǠȽɋǠȄƌ‫ژ‬uِ‫ژ‬°ƩǚȏƨƷȵ

BOARD OF GOVERNORS ¾ƷȵȂ‫ ژ‬-ɱȲǠȵǠȄǒ‫
ژي׎א׎א ژ‬Ȅưɲ‫ ژِ¾ ژ‬ǚƷȄً‫ ژ‬eȏǚȄ‫ ژ‬%ِ‫ ژ‬eȏǚȄȽȏȄً‫ژ‬ °ɲ‫ٮ‬äƷȄ‫ ژ‬hɓȏً‫ ژ‬%ƌɫǠư‫ ژ‬kȏȂƷɋً‫ ژ‬%ǠȂǠɋȵǠȏȽ‫ ژ‬°ƷȵȲƌȄȏȽً‫ژژژژژ‬ Oƌɲƌɋȏ‫ژ‬äƌȂƌȄƌ ¾ƷȵȂ‫ ژ‬-ɱȲǠȵǠȄǒ‫ ژي׏א׎א ژ‬uِ‫ ژ‬ȵǠƌȄ‫ ژ‬ǹƌǵƷً‫ ژ‬FȵƷư‫ ژ‬%ȏɓǒǹǠȽً‫ژ‬ ƌȵǹȏȽ‫ ژ‬-ِ‫ ژ‬eǠȂƷȄƷɼ‫ٮ‬GȏȂƷɼً‫¨ ژ‬ƌȂƌǹƌɋǚƌ‫ ژ‬uƌȵǠȂɓɋǚɓً‫ژژژژژژژژژژژ‬ -ȵǠǵ‫ ژ‬eƌȄ‫ ژ‬uƌȵǠȄǠȽȽƷȄً‫ ژ‬hɓȄǠȏ‫ ژ‬ÅƩǚǠɲƌȂƌ ¾ƷȵȂ‫ ژ‬-ɱȲǠȵǠȄǒ‫ ژيאא׎א ژ‬wǠǹȽ‫
ژ‬ȽƩǚƷȄƨȵɓƩǵً‫ژ‬ -ȵȄƷȽɋȏ‫ ژ‬ɓƌưȵȏȽ‫ٯ‬ÝƌȵǒƌȽً‫ ژ‬%ƌɫǠư‫ ژ‬°ِ‫ ژ‬-ƨƷȵɋً‫ ژ‬ÞǠǹǹǠƌȂ‫ ژ‬GȵȏȲȲً‫ژ‬ GȵƌƩƷ‫ ژ‬kƷɬǠȽً‫ ژ‬°ɋƷǑƌȄȏ‫ژ‬îƌȄƷȵȏ

EXECUTIVE STAFF Executive Director: Melissa
ِ‫ژ‬Russell Director, Governance & Associate Executive Director: Anne Marie Kelly Director, Finance & Accounting: Sunny Hwang Director, Information Technology & Services: Sumit Kacker Director, Marketing & Sales: Michelle Tubb Director, Membership Development: Eric Berkowitz

COMPUTER SOCIETY OFFICES Washington, D.C.: 2001 L St., Ste. 700, Washington, D.C. 20036-4928ٕ Phone: +1 202 371 0101ٕ Fax: +1 202 728 9614ٕ‫ژ‬ Email: ǚƷǹȲ‫ۮ‬ƩȏȂȲɓɋƷȵِȏȵǒ Los Alamitos: 10662 Los Vaqueros Cir., Los Alamitos, CA 90720ٕ‫ژ‬ Phone: +1 714 821 8380ٕ Email: [email protected]

u-u-¨°OU¥‫ژۯژ‬¥ÅkU
¾U‚w‫¨‚ژ‬%-¨°‫ژ‬ ¥ǚȏȄƷ‫ژٕבבבגژזוהژ׎׎זژ׏ڹژي‬Fƌɱ‫ژٕ׏גהגژ׏אזژג׏וژ׏ڹژي‬ -ȂƌǠǹ‫ژي‬ǚƷǹȲ‫ۮ‬ƩȏȂȲɓɋƷȵِȏȵǒ

IEEE BOARD OF DIRECTORS President: ¾ȏȽǚǠȏ‫ژ‬Fɓǵɓưƌ President-Elect: °ɓȽƌȄ‫ژ‬hِ‫ٹژ‬hƌɋǚɲ‫ژٺ‬kƌȄư Past President: eȏȽƸ‫ژ‬uِFِ‫ژ‬uȏɓȵƌ Secretary: Kathleen
ِ‫ژ‬Kramer Treasurer: Joseph V. Lillie Director & President, IEEE-USA: eǠȂ‫ژ‬ȏȄȵƌư‫ژ‬ Director & President, Standards Association: Robert S. Fish‫ژ‬ Director & VP, Educational Activities: °ɋƷȲǚƷȄ‫ژ‬¥ǚǠǹǹǠȲȽ‫ژ‬ Director & VP, Membership ‫ ۯ‬Geographic Activities:‫ژژژژژژژ‬ hɓǵDZǠȄ‫ژ‬ǚɓȄ Director & VP, Publication Services & Products: ¾ƌȲƌȄ‫ژ‬°ƌȵǵƌȵ‫ژ‬ Director & VP, Technical Activities: hƌɼɓǚǠȵȏ‫ژ‬hȏȽɓǒƷ

Column

From the Editor’s Desk Gerardo Con Diaz University of California, Davis

& I AM THRILLED to introduce our second issue of the year. We open in Italy during the 1950s, when the University of Pisa and Olivetti led the first Italian project to create an electronic calculator. Giovanni Cignoni and Fabio Gadducci dive deep into the University’s archives to reassess the emergence of computer science in Italy as a business and a scientific discipline. From Pisa we move to Binago, the small town north of Milan where an internationally renowned model maker named nin and Roberto Guatelli was born. Silvio He Simona Casonato examine Guatelli’s life and work to reflect on what it means to create, preserve, and trade replicas of ancient machines. Travelling east, we go to New Zealand to reassess the standard narratives about its first computer. Brian Carpenter shows that IBM was not, in fact, the first brand of computer to arrive at that country, as an ICT 1201 had already been in place for months by the time the New Zealand Department of Education acquired its first IBM 650. We end this world tour of computing among a global community of astronomers. Michael Scroggins and Bernadette Boscoe examine the history of FITS, the file format that astronomers have used since the 1980s to overcome incompatibilities between their operating systems. Dave Walden, our Anecdotes editor, has prepared two anecdotes for us. First is a very unusual one by Eileen Buckholtz that recounts the origins of Queens of Code—a project she is leading to preserve the histories and experiences of the women Digital Object Identifier 10.1109/MAHC.2020.2992095 Date of current version 29 May 2020.

April-June 2020

who worked in information technology at the National Security Agency from the 1960s to the 1980s. This Feature Anecdote gives us a glimpse of an exciting ongoing effort that has already recorded the histories of 75 queens of code. Zbigniew Stachniak reflects on the history of the MCM/70, one of the earliest computers for personal use to be mass manufactured. Stachniak has written extensively about this computer before, but new insights about its design and marketing emerged after the York University Museum in Toronto acquired additional prototypes of the machine. This anecdote reveals that we need to think about Toronto’s computing history more carefully when we account for the history of personal computing. The issue closes with an interview with Mike Williams, the 2007 IEEE Computer Society, and a book review of Meryl Alper’s book, Giving Voice. I also have exciting news to share with you: Thomas Haigh and Mark Priestley were awarded the 2019 Bernard S. Finn IEEE History Prize for “Colossus and Programmability” (IEEE Annals 40:4). Congratulations, Tom and Mark! To end this letter, I would like to express my gratitude. Even despite the chaos and uncertainty that surround us, the IEEE Annals is moving forward with its second issue of the year. This was possible because of the hard work and collaborative spirit of many members of our community. Our authors, department editors, article editors, and reviewers volunteered time from their unpredictable schedules to make this issue possible, and the IEEE staff was always ready to support us. Thank you.

Published by the IEEE Computer Society

1058-6180 ß 2020 IEEE

5

Article

Pisa, 1954–1961: Assessing Key Stages of a Seminal Italian Project Giovanni A. Cignoni HMR Project

Fabio Gadducci University of Pisa

Abstract—The last decade and a half have seen a renewed interest in the development of the IT industry in Italy and the role of the 1950s pioneers. The aim of the article is to retrace key stages of the first Italian project aimed at creating an electronic calculator, carried out by the University of Pisa in collaboration with Olivetti. The re-evaluation of the documents kept in the University’s archives, including the technical projects, has proved fruitful and has shed light on lesser-known aspects of a project that opened the way to the birth of computer science in the country as a business and as a scientific discipline.

& DESPITE THE ATTENTION paid early on to some of the protagonists, the Olivetti company and the figures of Adriano and Roberto Olivetti,1 research on the history of Italian computer science began rather late, the pivotal event being the 1991 conference organized by the Italian Association for Computer Science and Automatic Computing.2 However, the last decade and a half have been fruitful with new investigations, starting with the Milan meeting in 2004, which celebrated the fiftieth anniversary of the arrival of the first computer in Italy.3 It was followed by events focusing on the results of Rome4 and Pisa,5 thus covering the centers that introduced computer science in the country.6

Digital Object Identifier 10.1109/MAHC.2020.2978162 Date of publication 3 March 2020; date of current version 29 May 2020.

6

1058-6180 ß 2020 IEEE

All these events celebrated the pioneering Italian experiences of the 50s, usually with the participation of the protagonists of the time and often recording their memories. Current research is, however, going beyond these memories, providing a documentary context for these initiatives and pushing the exploration of archives. As far as Pisa is concerned, there are now repositories that make available many original documents,7 which are beginning to be evaluated as a whole.8 The focus of the article is on tracking down less explored aspects of the Pisa Electronic Calculator (CEP) project, run by the local university. The project started in 1954 with the ambitious goal of building an electronic computer from scratch, the first of its kind in Italy. The main partner was the Olivetti company, who used the experience to start its own line of commercial computers, the

Published by the IEEE Computer Society

IEEE Annals of the History of Computing

first of which to be marketed was the transistorized ELEA 9003 in 1959.9 Although main features of the history of the CEP project have already been outlined and its relevance for the development of computer science in Italy as a business and as a scientific discipline is by now well recognized,10 there are still some pivotal moments and milestones in the history of the project that need to be further investigated, in order to properly evaluate the outcomes of the project and better understand its impact on the emerging Italian computer industry. The paper will, therefore, cover three key stages of the CEP project. 1. First of all, we will discuss its birth, to adequately evaluate the role of Enrico Fermi in the process. His towering figure has overshadowed the group of scientists who actively pushed for the project, eclipsing the role of the Italian scientific community. 2. We will then consider the first machine built by the project, referred to as the MR (1957), and usually considered a prototype of minor importance. On the contrary, the careful analysis of the technical blueprints of Centre for the Study of Electronic Calculators (CSCE), the university institute managing the project, reveals a state-of-the-art machine, testifying to the diffusion of technological advances and their fast reception in the country. 3. Finally, we will discuss the difficulties of the second phase of the project, including the administrative transfer of CSCE from the University of Pisa to the National Research Council (CNR). We will analyze the impact of these difficulties on the construction of the final machine, revealing why, despite the building of the MR, the outcome of the project was in 1961 a computer that, in the words of an observer, was interesting but late.11

BIRTH OF THE PROJECT AND THE ROLE OF ENRICO FERMI An established tradition about the birth of the CEP project concerns the funding granted to the University of Pisa by a consortium of local public bodies, the University Interprovincial Consortium (CIU). It is well known, as confirmed by the memories of the protagonists, that the construction of a computer was not the CIU’s first choice.

April-June 2020

The funds provided by the municipalities and provinces of Pisa, Livorno, and Lucca should have been used for the construction of a synchrotron, designed by the Institute of Physics of Pisa. Eventually, the synchrotron was built in Frascati with funding provided by the municipality of Rome.12 Only later were the CIU funds diverted to the CEP project. It is, therefore, surprising and indicative of the partial nature of the Pisa archives that very few traces relating to the funding of the synchrotron survived. In the minutes of the Executive Committee of the CIU, the synchrotron is only mentioned on 20 May 1955,13 when this project had since long been abandoned and the decision to build a computer was taken. A summary14 mentions an unrecorded meeting on 20 March 1954, during which the commitment to the synchrotron was likely decided. The summary, however, is much later, as it mentions the choice of the splitting of funds between the electronic calculator and the mass spectrograph established at the 4 October 1954 meeting.15 The tradition of Enrico Fermi’s involvement in the diversion of the CIU funds is also established, and it includes an independent suggestion by the scientist to Pisa Rector Enrico Avanzi. In addition to the letters exchanged between Fermi and Avanzi in August 1954,16 a clearer view about Fermi’s role can be grasped from the correspondence between Marcello Conversi,17 director of the Physics Institute of Pisa, and Gilberto Bernardini,18 president of the National Institute of Nuclear Physics (INFN), with Mauro Picone, director of the National Institute of Applied Mathematics (INAC) in Rome. In fact, Fermi is often still remembered19 as the single voice pushing for the construction of an electronic computer, at the meeting of the International School of Physics in Varenna. On the contrary, while the location is correct, the letters tell of “a discussion lasting days” during which the possible uses of CIU funds were discussed “in an atmosphere of dispassionate objectivity and clarity.”20 The result was that the construction of an electronic computer was considered “by far the best choice among all the others,” as stated by Fermi in his endorsement to the Rector of Pisa.21 This letter was not an autonomous initiative of Fermi: as Conversi said, it was written “to my and

7

Article

prof. Bernardini’s request.”22 Avanzi replied to Fermi declaring “to be pleased that he had discussed with colleagues Conversi and Salvini the possibility of providing the University of Pisa with a scientific tool of national importance,”23 suggesting that the meeting had been considered by the community of Italian physicists as the ideal context to discuss the new allocation of funds, to present to the CIU a proposal that had strong and authoritative support from a large number of Italian scientists. However, the actual role of Fermi24 helps to understand the social context in which the CEP project was born and highlights the importance of the support of the scientist to unlock the hesitations of Pisa. It is indicative that in the October 4, 1954 meeting two important politicians, Pagni and Maccarrone (respectively, Mayor of Pisa and President of the Province of Pisa, members of the Board of Directors of the CIU), expressed their regrets: despite the recognition of the potential of an electronic machine, they stated that “the synchrotron exerted a greater influence on public opinion" and would have been “an easy subject of spectacular propaganda.”25 The episode testifies to the perception that Italian decision-makers had about these new computing devices:26 according to politicians, computers were not easily marketable with the public, while everything related to atomic energy was a source of fascination toward scientific and industrial progress.27 Fermi’s letter was read during the meeting of 4 October 195428 and in that same meeting, the choice of building a computer was ratified, although the formal steps for the start of the project had still to be completed. It was during the meeting of 13–14 January 1955 that the calculator began to be mentioned as “a suggestion from the late Professor Fermi.”29 The scientist had died on 28 November: the attribution could have been a sincere recognition motivated by the emotion for the recent death or, on the wave of the same feeling, a way to overcome the skepticism within the University. In fact, at the same meeting, the Faculty of Engineering still expressed concerns about the feasibility of the project. The correspondence between the Dean of Engineering Enrico Pistolesi and Avanzi, where the Rector has a strong position of support for the computer, is also remarkable in this regard. In a letter dated January 21,

8

Pistolesi argued that “a committee of the Faculty of Engineering should first examine the opportunities and convenience of proceeding with the design of the machine, a subject on which many colleagues have expressed concerns.”30 The letter also reveals some of the Faculty’s interests: using CIU funds for new buildings. Avanzi’s reply of January 31 was firm: he reiterated that, as decided in previous meetings, the project will start and its control will be entrusted to a university committee and renewed to Pistolesi “the request for appointing a representative of his Faculty.”31 Fermi’s support was also used to prevent the opposition of INAC. The Institute of Rome was the most advanced research center in Italy dedicated to computer science and was negotiating the purchase of a Ferranti computer.32 In a letter to Conversi, Aldo Ghizzetti reports the harsh words of the head of the INAC to the news of the CEP project “I deplore the Pisa initiative to build an electronic computing machine. . . I will oppose with all my strength the waste of money that would occur as a result of the approval of the Pisa initiative.”33 In their letters to Picone, together with a diplomatic deference (the CEP is declared “definitively second” to the Ferranti of INAC), Conversi and Bernardini used Fermi as an indisputable supporter of the project.34 The story had a happy ending:35 Picone’s position began to dissolve in December 195436 and INAC collaborated in the CEP project. On March 9 1955, the CSCE is finally established:37 the members of the Steering Committee are the physicist Conversi, the mathematician Alessandro Faedo, and the engineer Ugo Tiberio. Given the now general agreement on the project, in the preface, the choice of the computer is simply described as the result of a discussion “at the Physics Congress held in Varenna with foreign colleagues.” The words “Fermi’s suggestion” never again appear in the minutes. No similar statement is present either in the brochures published in 195938 or in the popular articles39 that appeared in those years. Fermi is not mentioned during the inauguration of the 1958/59 academic year when the Rector remembers the CSCE together with its first success, the construction of the MR.40 While the support of the scientist was likely decisive for the start of the project, it is a

IEEE Annals of the History of Computing

rhetorical exaggeration to identify him as the originator of the CEP project, and it overshadows the joint efforts of the scientific community. The facts are best summed up by quoting a 1958 internal note of the CSCE:41 the choice to build a computer was the result of “consultations that the professors of the University of Pisa had in Varenna in July 1954 with various physicists of international renown, among whom we remember, in particular, Enrico Fermi.” Nevertheless, identifying the CEP project as “Fermi’s last gift to Italy”42 has become a topos in the history of Italian computer science, strengthened in official ceremonies and even in the memories of witnesses.43

FIRST OUTCOME: THE MACCHINA RIDOTTA

The CSCE Steering Committee44 had a general control function, also managing relations with the academic and scientific world. An Executive Group carried out the study and design activities. Olivetti was mentioned since the first meetings,45 and the collaboration formalized in May 1956.46 Initially, the main members of the research group were Alfonso Caracciolo, Giuseppe Cecchini, Elio Fabri, and Sergio Sibani.47 At the beginning of the project Mario Tchou, Olivetti’s chief engineer, had the role of administrative manager,48 but he was later completely absorbed by the management of the Laboratory of Electronic Research (LRE), established by Olivetti in Pisa during the fall of 1955 and moved in the spring of 1956 to Barbaricina, in the town suburbs.49 While Caracciolo and Fabri were in charge of the logical and architectural design of the machine, Cecchini and Sibani designed the electronic implementation.50 The project staff grew steadily and in March 1958 it counted over thirty people, including both technical and administrative staff.51 The first year of CSCE was a period of study, during which the researchers acquired the necessary know-how. As an internal note points out,52 the Centre initially had some financial difficulties due to delays in the disbursement of funds. From the beginning of 1956, though, the activities were intense, and many feasibility experiments were carried out.53 At the end of this period, the CSCE achieved an important result: the completion of

April-June 2020

the detailed design of the prototypical Macchina Ridotta (“smaller machine,” in the following MR).54 The project was signed by Caracciolo, Cecchini, Fabri, and Sibani, with thanks to Menotto Baldeschi, Giovan Battista Gerace, and Vladimiro Sabbadini.55 Most of the surviving technical information on the MR is made up of drawings.56 It is worth noting that these documents refer to the first detailed project, which is substantially different from the machine that was actually built. Another report is dated April 1957 and precedes the completion of the machine by a few months.57 It describes in detail the numerous changes since the 1956 project, mainly due to the need to increase usability and the possibility of being connected to multiple input and output devices. The note has been written by the logic-mathematical section of CSCE: Caracciolo and Fabri. In fact, most of the modifications, even if substantial, were of an architectural nature and did not require changes to the electronic design of the individual components. The report mentions some additional blueprints that are no longer present either in the archives of the University of Pisa or in those of the CNR. Only one has been recovered, but it is essential to the understanding of the “smaller machine”: the new version of the general scheme.58 The MR was ready in July 1957, as witnessed by a letter that Conversi addressed to many colleagues at Italian universities.59 In addition to giving news of the milestone, the scientist discussed the possible uses of the MR for computing services, making it available to colleagues and starting to look for real case studies in order to test capabilities and performance. The writing of a manual, in fact, testifies the presence of users for the machine.60 In 1958, four researchers from four different INFN centers joined CSCE, with the aim of acquiring skills in computer programming. The group included two , which women, Elisabetta Abate and Marisa Rome 61 is noteworthy for the period. In particular, Abate drafted the user manual of the machine: the practical and succinct style of the document shows that it was written for users, possibly external to CSCE, developing algorithms in machine language.62 The document lacked any instruction on how to operate the MR: a task for the same designers and engineers who built it.

9

Article

A careful examination of the surviving designs reveals interesting details, such as the dating of the drawings, sometimes contrary to expectations. Considering, for example, the blueprints of the binary adder, the logical network (the specification63) is dated 11 July 1956, while part of the electronic circuit (its implementation) is earlier, dating to 20 June (sum circuit64) and 16 June (carry circuit65). Moreover, in the latter blueprints, the notation of the logical gates is reversed: in one case the white triangles are AND and the black ones OR gates, in the other the opposite. Moreover, the electronic circuits have a small defect that prevents them from working properly. All these problems were obviously solved, but they left no trace in the surviving designs. Anecdotally, they confirm that, as far as design documentation is concerned, the bad habits of computer scientists are deeply rooted. Finally, the analysis of the blueprints allows us to compare the MR with the technology of the time and to understand how the Pisa researchers built their know-how. The logical network of the adder is built according to a solution that reduces the number of levels of the logic gates and, therefore, the calculation time. This solution appeared in an article describing the arithmetic unit of the IBM 701.66 The comparison with CSCE projects leaves no doubt about the sources that inspired the designers,67 even if the relationship was never declared since neither the article nor the IBM origin of the solution is mentioned in the surviving CSCE documents: a confirmation of the need to verify the technical details in order to properly asses the spread of the technological advances.

ACCOMPLISHMENTS OF THE MACCHINA RIDOTTA Since the beginning of 1958, the MR was used for research purposes. There are several reports about the presence of users outside CSCE. The first computation service was requested by the Institute of Mineralogy of the University of Pisa for a work on crystallography. The project was completed in April 1958, and the execution of the program took 80 minutes.68 Other services included a 60 hours computation, in multiple sessions, part of a research on radio frequencies

10

in the ionosphere of the University of Rome. By the end of 1958, the computer had performed about 150 hours of external services for a value that “can be estimated at 8 million lire,” as stated in an internal report.69 Although the birth of CSCE had already been mentioned in 1956 in the newsletter of the U.S. Navy Research Centre,70 the publication of the results obtained by the MR aroused greater interest. At the end of his mandate, the Rector Avanzi received a request from the scientific  of the American Embassy in Rome to attache visit CSCE: the meeting was held on 30 October 1959.71 In order to have a better perspective on the work of CSCE Executive Group, it is worthwhile to highlight some of the key features of the architectural design of the "smaller machine,” namely 





parallel bit processing: When the CEP project began most computers were “serial,” i.e., the bits of a memory word were processed one at a time in successive clock cycles; the MR was instead “parallel,” processing all the bits of a word in a single clock cycle; ferrite cores memory: Instead of one of the memory technologies of the early 1950s, such as magnetic drums, acoustic delay lines or Williams tubes, CSCE researchers chose ferrite cores, adopting an emerging technology that will be dominant for the following two decades; micro-programmed control: The MR used the micro-programming solution proposed by Wilkes, even if it used a less sophisticated technology (removable diodes instead of ferrite rods) than the one used on EDSAC 272, and the implementation was made easier by the small number of instructions (32), all executed in two microinstructions (fetch/execute, without cycles).

While the first two features were standard for machines designed by mid-Fifties and based on current technology, the MR appears to be one of the earliest microprogrammed computers.73 However, what is most enlightening about the international dimension of the CSCE researchers is that none of these features occurred on the other two computers present in Italy since 1955,74 even if

IEEE Annals of the History of Computing

these machines, geographically close and working within public research facilities, could have been easily accessed and studied and, thus, possibly used as models for the Pisa project. The MR was also a remarkable achievement in terms of performance. With careful tuning, the execution time of the instructions was reduced by 30% compared to initial estimates, resulting in a clock cycle of 4 or 8 ms depending on the current microinstruction. Conversi declared the resulting performance of more than 60 000 instructions per second “superior to all machines on the market, including the IBM 704 located in Paris.”75 It should be noted that the “superiority” only affected speed and was influenced by the different sets of instructions. The 704 had more memory and peripherals, not to mention the Fortran compiler. However, being able to challenge the performance of the IBM machine on the most direct benchmark remains a result to be proud of. For the sake of fairness, we must point out an obvious defect of MR: although it considered the use of subroutines, it had no specific instructions for jumping to them. At the time, the Wheeler Jump used in EDSAC was a well-known solution.76 It might have been implemented on the MR with minimal hardware changes. The Pisa researchers did not implement it, although in other respects (micro-programming) they had adopted the results of the Cambridge group. It may, therefore, sound surprising that until the conference in Pisa in 2011,77 the technological relevance of the MR was not properly recognized.78 However, there seems to have been some reasonable causes that might have contributed to this forgetfulness, namely 





among the four designers, only Caracciolo remained at CSCE; Fabri and Sibani left CSCE in 1959,79 Fabri returning to astrophysics, Sibani instead to Olivetti, as Cecchini will do in 1961; there was no surviving physical evidence for the MR, as it was completely dismantled and its electronic materials reused for the construction of the CEP; there was no official inauguration of MR and, due to political tensions between the Italian Universities and the national Government, not even the traditional inauguration ceremony of

April-June 2020



the 1957/58 academic year took place; 80 the construction of the MR will only be remembered in the prolusion of the following year81 when it had already been dismantled; the CSCE work plan for the years 1956/57 did not mention explicitly a first computer, but “the heart of the machine, i.e., the entire machine with the exception of external devices: the magnetic drum and the fast input/ output devices.”82

We believe that the crucial reason is probably the characterization of the ‘‘smaller machine” as the “core” of the CEP, a description that later also appeared in several reports. Since the official documents suggested the interpretation of the MR as an important step, yet an incomplete part of the final computer, the statement was taken for granted by later historians and its importance in the development of the CEP project was then underestimated. On the contrary, a careful analysis of the technical blueprints shows that the two machines were very different (see also Table 1). For example, the memory, one of the components that would have been easier to reuse with very few modifications, was completely redesigned: it consisted of 32  32 single-sided planes in the first computer and 64  64 double-sided planes in the second. And this is just the most straightforward example: from the microprogrammed control to the electronics of the adder, the differences between the two machines were numerous and significant. A few authors cite the MR and provide some details,83 but rather than the machine that was actually built in 1957 they describe the 1956 design, possibly because there are more surviving copies of the corresponding report.84 However, that was just a first draft, and it described a different and simpler computer. By comparing the two designs it is possible to appreciate the differences: the presence of additional devices (a second teletypewriter, the tape puncher, a second tape reader), the flexible management of the I/O operations, the easy boot of the system software with a sort of “direct memory access,” and the mechanism of hot breakpoints for debugging. Other improvements in terms of usability are reflected in the design of the manual control

11

Article

Table 1. Technical data of the two computers built by the university of pisa. Smaller Machine, 1957–58

CEP, 1961–69

Word

18 bit

36 bit

Memory

1024 words, ferrite core, 3232 single-sided planes

4096 words (later 8192), ferrite core, 6464 double-sided planes

Logic implementation

Diode-resistor logic, triode inverters

Diode-resistor logic, triode inverters, transistor inverters in a few cases

Micro-programmed control

Diode matrix, fetch or execute micro-instructions

Ferrite rods matrix, pseudo instructions, conditional and cyclic micro-instructions

Number of instructions

32

128

Fixed point additions per sec

62 500

67 000

Floating point additions per sec

Not available as machine instruction

10 400

Support for subroutines and array operations

Address substitution

Double indirection using index cells

I/O devices

1 teletypewriter Olivetti T2CN 1 teletypewriter Olivetti T2CN-PF with tape puncher 1 tape reader Olivetti T2TA10 1 fast tape reader Ferranti TR5

1 teletypewriter Olivetti T2CN 2 fast tape readers Ferranti TR6 2 fast tape punchers Teletype LMU6 1 line printer Bull 1 magnetic drum (32768 36 bit words) 6 magnetic tape readers (later)

System software

Basic arithmetic sub-routines, simple program loader

220 math and utility sub-routines, symbolic assembler, FORTRAN compiler (later)

panel and in the visual feedback on the last read/written memory word and on the value of the program counter. The comparison between the two versions of the design witnesses the remarkable work done by the CSCE researchers between 1956 and 1957. It is a process that shows the importance of MR in the CEP project and, in general, in the formation of the first Italian computer scientists. Another unfortunate point is that very little of the photographic documentation of the MR survives. The most significant is the image in Figure 1, which shows the three racks that made up the machine: from left to right, the arithmetic-logic unit, the micro-programmed control, and the memory. In the background, below the window, you can see the back of the control panel. At the far right of the image, the second teletypewriter with a tape puncher appears in the foreground.

DIFFICULT YEARS The success of the MR closed the first phase of the project. Unfortunately, the work toward

12

the following objective, the construction of the CEP, took place in a difficult period for CSCE. The first problem was transistor technology, whose production had boomed since mid-1958.85 The possible change in the choice of hardware was discussed at length within the Executive Group,86 in the knowledge that the original financing was about to end. And although the hopes of new funds were rosy and lasted for some time,87 these did not materialize, although the CSCE had shown its ability to keep pace with the most advanced international projects. The conversion to transistors was, therefore, not possible, even if they were used in the CEP microprogrammed control, which instead of the diodes used ferrite rods to implement the ROM matrix (a variant of the EDSAC 2 solution). In addition to the hardware costs, software development was another problem. A substantial effort to build software libraries was planned in time.88 However, again in February 1961, while the machine’s hardware was missing only a few tests, there were concerns about the delay in the development of the system software. To work on CEP software before the completion

IEEE Annals of the History of Computing

Figure 1. MR at the Department of Physics.

of its hardware, collaboration was set up with INAC to develop a simulator using the Institute’s Ferranti computer.89 The CEP was operational in spring 1961, a year later than planned.90 Despite its limitations, the CEP was a relevant machine. The U.S. observers91 recognized its main features, including speed, microprogramming, and indirection mechanisms to support subroutines. However, as Blachman bluntly points out, the CEP was late. On the contrary, while the CSCE was suffering from a lack of funding, Olivetti had invested and accelerated the pace. The LRE of Pisa, almost simultaneously with CSCE, developed the “zero machine,” the first Olivetti computer (sometimes also called Elea 9001). A second prototype, called Elea 9002, was completed in 1959, with the redesign of transistor electronics; it was installed at the Olivetti headquarters in Milan and inaugurated on November 8, 1959. The commercial product, called Elea 9003, was announced at the Milan Fair in April 1959, and the first machines were delivered to customers in 1960.92

PREVIEW OF THE CALCOLATRICE ELETTRONICA PISANA The CEP was inaugurated on 13 November 1961 in the presence of President Giovanni Gronchi (Figure 2). In fact, as we read in a letter

April-June 2020

from Conversi to Faedo concerning a call for a project,93 CSCE tried to move up the ceremony to give “a small resonance in time to get some

Figure 2. Protagonists of the preview: From left-toright: Giovan Battista Gerace, Alfonso Caracciolo, the President of the Republic Giovanni Gronchi, Marcello Conversi.

13

Article

Figure 3. A picture distributed to the press at the preview. The machine is shown open during maintenance to highlight its complexity. Two CSCE employees are in view: At the oscilloscope Luigi Pistelli, at the console Luciano Azzarelli.

additional funding.” The President’s agenda, however, delayed the ceremony until the autumn.94 Nevertheless, the opening ceremony95 was a big event for the city of Pisa. The University’s press release96 and the newspaper articles appearing in those days97 retrace in some detail the history of the project (Figure 3). In the memory of the protagonists, however, the inauguration was not interrupting the bleak period: the goal had been achieved but the funding was over and the future was more uncertain than ever. In December 1961, a new statute of the CSCE was drawn up, with the hope of receiving the approval of the CNR. In July 1962, an agreement was signed that started the transition process: the CSCE became a “CNR Institute of the University of Pisa.”98 The CNR was now responsible for the staff, while the University was responsible for the structures. The management was entrusted to a body that saw an equal share of representatives of the University and the CNR. The convention lasted until 1968 when the CSCE became the Institute for Information Processing of CNR.99 The Center achieved the stability that will allow further developments of the CEP (e.g.,

14

doubling the memory and adding magnetic tape drives) and to use it both to do research on the software side (such as the Fortran compiler) and to offer computing services to the outside. The CEP will remain in operation for a decade, working day and night. In 1966, the machine is still accounted on average for more than three hundred machine hours per month, with peaks of almost five hundred: as in March, when the CEP worked for 496 hours, 170 of which for users outside the CSCE.100 In the Executive Group, there will be new arrivals, while some of the early protagonists will depart soon. We already mentioned the fate of the four designers of the MR: Fabri will go on in his work as a physicist, and Conversi will follow suit shortly after the completion of the CEP. Cecchini and Sibani will resume work in Olivetti as engineers, then moving to the newly formed Olivetti-General Electric joint venture in 1964.101 Caracciolo and Gerace will play a pivotal role in the development of Italian research in Computer Science, and together with by then Rector Faedo, they will help to create the Degree in Computer Science, the first of its kind in Italy, established by the University of Pisa in 1969.102 Indeed, the

IEEE Annals of the History of Computing

influence of CEP on the production of Italian mainframes will be minimal, and their production will anyhow cease by the end of the 60s, when General Electric sells its Italian subsidiary to Honeywell. The lasting heritage of the CEP project is the human capital Pisa will produce, attracting future researchers and becoming a hub of research centers by the like of HP, IBM, and Olivetti, as well as hosting most of the institutes devoted to Informatics of the CNR, thus transforming Pisa into the “Italian town of Computer Science” in the following decades.

CONCLUDING REMARKS The article has investigated the Pisa archives with the aim of shed further light on less-known aspects of the CEP project, a seminal enterprise for the development of computer science in Italy. Besides the layers of complexity that are, thus, added to the history of the project, each of these facets is in itself a litmus test for a different aspect of the research activity and practice. By eliminating the almost hagiographic aspects, Enrico Fermi’s involvement at the beginning of the CEP project helps to outline the role of the institutional actors: the firm commitment of Rector Avanzi, the conflicts in the Pisa academic community, and the initial doubts of the political representatives. In the end, the story is a paradigmatic witness of the need for public endorsement by key figures in innovative projects, as well as the risk its overestimation poses to historical enquiry. The recollection of the MR characteristics testifies to the importance of a careful analysis of technical issues, in order to faithfully reconstruct the results of a research project and its connections with similar enterprises, adequately assessing its value. Furthermore, the results of the analysis can be used for educational purposes, as evidenced by the use of the original designs for the construction of the MR simulators103 and the working replica of the 6-bit adder.104 Finally, the difficulty of building the 1961 CEP is an exemplary case of the need for continuous funding of technological research. A belief that the Italian government at the time did not support, since the funding was most often given

April-June 2020

for the acquisition of machines, and the first national plans concerning computer science will get funded only at the end of the 1970s. But most importantly, it is a lesson that still holds today as it did in the 1950s.

NOTA BENE: REFERENCES TO DOCUMENTS The [n] item added to the references of archival documents is a pointer to the list on the https://www.progettohmr.it/IEEEAnnals webpage, where digital copies of these documents have been made available to the reader. Often not cataloged or difficult to find, these documents come from the General Archive of the University of Pisa, the Library of the Institute of Science and Information Technology of the CNR of Pisa, the Archive of the Institute for Computer Applications of the CNR of Rome, and the private archive of Elio Fabri.

& REFERENCES AND ENDNOTES 1. The bibliography on the life of Adriano Olivetti and his period at the helm of the family firm is quite large: for a biography of the entrepreneur and politician see V. Ochetto, Adriano Olivetti. La biografia. Edizioni di  , 2013, while for the fate of Olivetti’s Comunita electronics division see at least G. Gemelli, Normalizzare l’innovazione. Le vicende dell’elettronica e dell’informatica da Adriano a Roberto Olivetti. Fondazione Adriano Olivetti, 2014. Roberto was Adriano’s son, and later on he took care of the informatics division: see D. Olivetti (ed.), Roberto Olivetti. Fondazione Adriano Olivetti, 2003. A recent biography of the main designer of Olivetti mainframes is G. Parolini, Mario Tchou. Ricerca e sviluppo per l’elettronica Olivetti. EGEA, 2015. For an English summary of the early Olivetti computers, even if mainly focused on the industrial/economical issues, see G. Parolini, “Olivetti elea 9003: Between scientific research and computer business,” in J. Impagliazzo (ed.), Proc. 3rd IFIP Conf. History Comput. Educ., IFIP AICT 269. Springer, 2008, pp. 50–64; while for a more encompassing, popular piece see the very recent E. Mori, “The italian computer: Olivetti’s ELEA 9003 was a study in elegant, ergonomic design,” IEEE Spectrum, June issue, 2019.

15

Article

2. The conference was held on September 10–12, 1991,

therein, such as G. Rao, “Mario Tchou e l’Olivetti ELEA

(ed.), Convegno internazionale sulla storia e preistoria

9003,” PRISTEM/Storia, vol. 12–13, pp. 85–119, 2006,

del calcolo automatico e dell’informatica. AICA, 1991.

it is worthwhile to check www.elea9003.it for a trove of

Another early collection of essays is VVAA, La cultura

images and documents about the last working ELEA

informatica in Italia: Riflessioni e testimonianze sulle origini, 1950–1970. Bollati Boringhieri, 1993. 3. See the proceedings of the event in L. Dadda (ed.), La nascita dell’informatica in Italia. Polipress, 2006. 4. A conference has been held in 2015 for the 60 years of

9003. 10. For an English-language outline of the history of the project, we refer to the surveys De Marco et al. 1999 and Bonfanti 2012. Concerning the impact of the project on the development of IT research in Italy we

the Ferranti acquisition (see www.cnr.it/it/evento/14303),

just note that most IT institutes of CNR are in Pisa and

but no proceedings so far appeared. For a survey on

have their roots in CSCE and that the University of Pisa

early computers in Rome see, e.g., P. Nastasi, “Picone, il

organised the first Italian degree in computer science

calcolo automatico e FINAC: una storia lunga 30 anni,”

in 1969: see, e.g., C. Montani et al., “Il CNR dopo la

PRISTEM/Storia, vol. 12–13, 2006, pp. 121–171, as

CEP,” Cignoni, Gadducci 2013, pp. 41–66.

well as Picone’s correspondence in A. Guerraggio,

11. N.M. Blachman, “The state of digital computer

M. Mattaliano, and P. Nastasi (eds.), La lunga marcia di  Mauro Picone, Quaderni PRISTEM n. 15. Universita Bocconi, 2010. 5. There were two conferences devoted to the Pisa results,

technology in Europe,” Commun. ACM, vol. 6, no. 6, pp. 256–265, 1961. 12. For an English-language history of the Laboratories of Frascati, see G. Salvini, “The electron synchrotron and

held in 2009 and 2011. The proceedings were

the birth of the national laboratories of Frascati,” in

respectively collected in M. Vanneschi (ed.), La CEP:

M. De Maria, M. Grilli, and F. Sebastiani (eds.), The

Storia, scienza e umanita dell’avventura informatica

Restructuring of Physical Sciences in Europe and the

pisana. Felici, 2009 and in G. A. Cignoni and

United States. Singapore: World Scientific, 1989,

F. Gadducci (eds.), La CEP Prima Della CEP: Storia

pp. 532–547.

Dell’Informatica. Pisa, Italy: Pisa Univ. Press, 2013. See also www.cep.cnr.it and cep50.di.unipi.it, respectively. 6. Among the books offering a general view of the period see the monographic issue VVAA, “50 anni di informatica in Italia,” PRISTEM/Storia, vol. 12–13, 2006, as well as the conference proceedings mentioned above. For English surveys of the first Italian projects see

13. Minutes of the CIU Executive Committee, meeting of May 20, 1955, pp. 59 and 73, AUniPi [1]. 14. Summary note about the funding from CIU, no certain date, but not earlier than October 1954, AUniPi [2]. 15. Minutes of the meeting of October 4 at the Dean House, 1954, AUniPi [3]. 16. E. Fermi, letter to E. Avanzi, August 11, 1954,

G. De Marco et al., “The early computers of italy,” IEEE

typewritten transcript, AUniPi [4] and E. Avanzi,

Ann. History Comput., vol. 21, no. 4, pp. 28–36, 1999,

letter to E. Fermi, August 24, 1954, archive copy,

and C. Bonfanti, “Information technology in Italy: The origins and the early years (1954–1965),” in A. Tatnall (ed.), Reflections on the History of Computing, IFIP AICT 387. New York, NY, USA: Springer, 2012, pp. 320–347. 7. See, in particular, www.progettohmr.it/ Documentazione/Archivio and csce.isti.cnr.it. 8. An early exploration is G. De Marco’s Master thesis “La Calcolatrice Elettronica Pisana: le origini dell’informatica in

16

9. Besides Parolini 2008, 2015, and the references

and the (pre-) proceedings appear as C. Bonfanti

AUniPi [5]. 17. M. Conversi, letter to M. Picone, October 12, 1954, AIAC [6]. 18. G. Bernardini, letter to M. Picone, November 11, 1954, AIAC [7]. 19. See passim. For a recent entry, the 2015 exhibition devoted to the scientist, witnessed in www. mostrafermi.it/navigator_4.html

Italia,” discussed in May 1996, which has been extensively

20. See Bernardini 1954 [7].

used by later historians. Among the recent additions,

21. See Fermi 1954 [4]. The original has been lost, while

see T. Paladini, “La macchina «CEP», come noi ci siamo

a typewritten copy survives thanks to the care of a

abituati a chiamarla. L’avventura pisana nell’indagine

secretary: see F. Denoth, “I primi calcolatori: La CEP

storiografica delle carte d’archivio,” in Vanneschi 2009,

pisana,” PRISTEM/Storia, vol. 12–13, pp. 127–140,

pp. 9–34 and G. A. Cignoni, F. Gadducci, and

2006.

D. Ronco, “I documenti raccontano le storie della CEP,”

22. See Conversi 1954 [6].

in Cignoni, Gadducci 2013, pp. 119–146.

23. See Avanzi 1954 [5].

IEEE Annals of the History of Computing

24. The scientist was keenly aware of the importance of

40. E. Avanzi, “Inaugurazione dell’a.a. 1958/59, 15

computers for research and industry: see, e.g., the

novembre 1958,” in Annuario dell’Universita degli

concise remarks in Bonfanti 2012.

Studi di Pisa, a.a. 1958/59, 1960. A transcript is

25. See Minutes 1954 [3].

available at www.sba.unipi.it/it/risorse/archivio-

26. M. Zane, “Il percorso italiano verso l’informatizzazione,”

fotografico/eventi-in-archivio/1958-inaugurazione-

Altronovecento, vol. 3, pp. 5:1–5:29, 2000. 27. Even if the support of both persons in favour of the machine will be strong. At a meeting held in November 25, during a discussion about the use of Province funds in support of the University, Maccarrone states that “it is

aa-1958-1959 41. Probably written by Conversi: “Notizie  di Pisa,” concernenti il CSCE dell’Universita internal draft, March 26, 1958, AUniPi [16]. 42. The sentence is attributed to Giulio Racah in a 1958

saddening that some washed-up relics think that it

speech in Pisa and often reported, for a recent

would be better to build classrooms instead of an

example, see E. Visentini, “Gli anni ‘70 e la Scuola

electronic brain,” with a clear reference to the Faculty

Normale,” Archeologia e Calcolatori, no. 20,

of Engineering: see next paragraph. 28. See Minutes 1954 [3]. 29. See Minutes of the meeting of January 13–14 at the Institute of Physics, 1955, AUniPi [8]. 30. E. Pistolesi, letter to E. Avanzi, January 2, 1955, AUniPi [9]. 31. E. Avanzi, letter to E. Pistolesi, January 31, 1955, AUniPi [10]. 32. A. Celli and M. Mattaliano, “Mauro Picone e i primi

pp. 11–15, 2009. 43. G. Salvini, “Enrico Fermi il maestro sperimentale e teorico del secolo ora trascorso. Alcuni personali ricordi,” Il Nuovo Saggiatore, vol. 17, no. 5–6, pp. 20–23, 2001. 44. According to document authorship, most of the job was actually carried out by Conversi. 45. See Minutes 1954 [3] and Minutes 1955 [8]. 46. Agreement between University of Pisa and Ing.

progetti per un calcolatore italiano,” in Cignoni,

C. Olivetti & C. SpA, May 7, 1956, AUniPi [17].

Gadducci 2013, pp. 3–19. The INAC Ferranti Mk1 is also

See Paladini 2009 for some of the doubts voiced

mentioned in De Marco et al. 1999 and Bonfanti 2012.

by members of the University of Pisa on the

33. A. Ghizzetti, letter to M. Conversi, October 26, 1954, in AIAC [11].The reference to funding is pivotal: Picone was then a major figure in the international

agreement with Olivetti, and the possible reasons for the late signing. 47. Caracciolo, Fabri, and Sibani came from the

scientific community, and had since long tried to

Institute of Physics in Rome and had been

raise government funds for the building of a

contacted in late 1954 by Conversi. They were

computer, as well as to establish partnership with

present since the meeting of January 1955 (see

Olivetti: see Celli, Mattaliano 2013.

Minutes 1955 [8]) during which Caracciolo

34. See Conversi 1954 [6], and Bernardini 1954 [7], respectively. 35. The reconciliatory words, together with a survey of

presented a comprehensive report on the state of the art about computers around the world. Cecchini was an Olivetti engineer. Formally, Caracciolo was

the comments by the Italian scientific community to

part of CSCE since its inception, Fabri and Cecchini

Conversi’s announcements of the CEP project, are

since June 1955, Sibani since September 1955;

reported in Paladini 2009.

Caracciolo and Fabri as employees of the

36. M. Picone, letter to G. Bernardini, December 30, 1954, AIAC [12]. 37. Minutes of the meeting of March 9 at the Dean House, 1955, AUniPi [13]. 38. “Informazioni generali sul Centro di Studi sulle Calcolatrici Elettroniche,” brochure, 1959, AUniPi [14]

University, Cecchini and Sibani as employees of Olivetti, “on lease” until the agreement (see Agreement 1956 [17]) was signed. See the CSCE staff listed at the end of Draft 1958 [16]. 48. See again the CSCE staff listed at the end of Draft 1958 [16].

and “General information concerning the

49. See Parolini 2008.

Center of Research on Electronic Computers (CSCE),”

50. A memory by Fabri recalls in details the skills,

CSCE Internal Report 1–47, 1959, AUniPi [15]. 39. M. Conversi, “Il Centro di studi sulle calcolatrici  di Pisa,” La Ricerca elettroniche dell’Universita Scientifica, vol. 1, no. 1, pp. 59–66, 1961.

April-June 2020

aptitudes, and inclinations of the members of the group: E. Fabri, “Gli inizi, il CSCE e la Macchina Ridotta,” in Cignoni, Gadducci 2013, pp. 21–29. 51. As it is also documented in Draft 1958 [16].

17

Article

 del CSCE dal 23-12-1955 al 52. “Relazione sulle attivita 31-07-1956,” internal document, AUniPi [18]. 53. As reported in Document 1956 [18], the experiments regarded the ferrite core memory (by Cecchini and Filippazzi), the design of the logical networks (Sibani and Galletti), the ferrite core memory (Cecchini and Gerace), the realization of the registers (Sibani), the construction and testing of a 6 bit adder (Sibani), and the design of the chassis for the assembly of the components. The presence of Filippazzi and Galletti

programmed by some researchers of the Chemical Institute. 63. Tecnical drawing “AD/L/1,” AISTI [25] (part of Caracciolo et al. 1956 [19]). 64. Tecnical drawing “AD/Ed/1,” AISTI [26] (part of Caracciolo et al. 1956 [19]). 65. Tecnical drawing “AD/Ed/2,” AISTI [27] (part of Caracciolo et al. 1956 [19]). 66. H.D. Ross, “The arithmetic element of the IBM type 701

witness the relationships with the LRE. The work on the

computer,” Proc. IRE, vol. 41, no. 10, pp. 1287–1294,

memory is the first contribution by Gerace to the project.

1953.

He arrived at CSCE in December 1955, later becoming a leader of the project and playing a pivotal role in the development of Computer Science in Pisa. See, e.g., the

67. Indeed, the designs of the logical adders do coincide. Compare Figure 6 in Ross’ article and Caracciolo et al. 1956 [25].

biographic entry in www.treccani.it/enciclopedia/

68. M. Conversi, letter to E. Avanzi, April 3, 1958, AUniPi [28].

giovan-battista-gerace_(Dizionario-Biografico).

69. “Sommario dei lavori svolti presso il CSCE nel 1958 e

54. A. Caracciolo, G. Cecchini, E. Fabri, and S. Sibani,

indicazione dei programmi di lavoro per il 1959,” internal

“Progetto dettagliato di una prima calcolatrice

document, November 13, 1958, AUniPi [29]. Also

elettronica (Macchina Ridotta),” CSCE Internal Report

interesting for its software solution is E. Abate and

1–26, 1956, AISTI [19].

E. Fabri, “Use of an Electronic Computer for the

55. Baldeschi was the draftsman, Sabbadini an Olivetti engineer at CSCE since July 1956. 56. Technical drawings folder, addendum to CSCE Internal

Construction of Exact Eigenfunctions of Orbital Angular Momentum in L-S Coupling,” Il Nuovo Cimento, vol. XIV, no. 1, pp. 29–47, 1959. The program made use

Report 1–26, 1956, AISTI [20], A. Caracciolo and E. Fabri,

of symbolic computation: a witness of the versatility of

“Complementi e variazioni al progetto logico dettagliato

the MR and of the research about programming

della Macchina Ridotta,” CSCE Internal Report 1–36, Apr. 26, 1957, AISTI [21], and Technical drawing “MR/S/2,” addendum to CSCE Internal Report 1–36, 1957, AFabri [22]. It is worth noting here that the numbering of the CSCE Internal Reports is the result of a later cataloguing: Initially the CSCE had no organized archive and documents (see Caracciolo et al. 1956 [19]) lack even the date. After a comparison with other reports, the earliest report (see Caracciolo et al. 1956 [19]) should be dated at the end of July 1956 and it represents the commentary to the massive collection of blueprints (see Caracciolo, et al. 1956 [20]) that completes the first design of the MR. 57. See Caracciolo, Fabri 1957 [21]. 58. Once more thanks to Fabri: see Caracciolo, Fabri 1957 [22]. 59. M. Conversi, newsletter to colleagues, July 24, 1957, AUniPi [23]. 60. E. Abate, “Prescrizioni fondamentali per l’uso della

techniques. 70. An., “Computer overseas - Centro Studi Calcolatrici Elettroniche (CSCE), Pisa, Italy,” Digital Computer Newsletter, vol. 8, no. 4, p. 15, 1956. 71. E. Avanzi, letter to W. Ramberg, October 26 ottobre 1959, AUniPi [30]. 72. M.V. Wilkes, “EDSAC 2,” IEEE Ann. History Comput., vol. 14, no. 4, pp. 49–56, 1992. 73. MR is roughly contemporary to EDSAC 2, “The first computer to have a micro-programmed control unit,” as stated in Wilkes 1992. 74. That is, the USA CRC102 at the Milan Polytechnic and the British Ferranti Mk1 at INAC. 75. Minutes of the CSCE Steering Committee meeting, Apr. 16, 1958, AUniPi, p. 3 [31] and Handwritten draft of the same minutes, AUniPi, p. 6 [32]. The reference is to the machine installed in 1957 in the IBM French headquarters assumed by the CSCE as a benchmark.

Macchina Ridotta,” CSCE Internal Report 1–38, Mar. 1,

Anecdotally, in the minutes the computer is cited as

1958, CSCE, AISTI [24].

“Fote IBM”: The typist misinterpreted the “704 IBM” on

61. Indeed, at the time no other woman researcher was included in the Pisa group, and from the surviving

18

62. As recalled in Fabri 2013, in one case, the MR was

the handwritten draft. 76. M.V. Wilkes, D.J. Wheeler, and S. Gill, The Preparation

documents the same appears to be for the Olivetti and

of Programs for an Electronic Digital Computer.

INAC teams in the mid-1950s.

Reading, MA, USA: Addison-Wesley, 1951.

IEEE Annals of the History of Computing

77. See Cignoni, Gadducci, Ronco 2013. 78. For example, only a few lines are devoted to it in De

91. See I.L. Auerbach, “European electronic data processing - A report on the industry and the state of

Marco et al. 1999, and it disappears altogether in

the art,” Proc. IRE, vol. 49, no. 1, pp. 330–348, 1961,

Paladini 2009. In fact, in the latter paper, it is stated that

Blachman 1961, and J. L. F. De Kerf, “A survey of new

in the history of the CEP project “the years 1957 and

West-European digital computers,” Comput. Autom.,

1958 are not characterized by relevant episodes.” 79. “Personale del CSCE,” internal document, Mar. 12, 1959, AUniPi [33]. 80. Communication of the University Senate, Nov. 1957, AUniPi [34].

vol. 12, pp. 27–28, 1963. 92. For a survey of the ELEA series see Parolini 2009 and Bonfanti 2012, as well as Mori 2019. 93. M. Conversi, letter to A. Faedo, March 15, 1961, AUniPi [39].

81. Avanzi 1960.

94. Paladini 2009.

 del CSCE al 22 dicembre 1955,” 82. “Relazione sulle attivita

95. “Programma di massima per la cerimonia inaugurale

addendum to the minutes of the meeting of December 22 at the Dean House, 1955, AUniPi [35]. 83. See, e.g., P. Maestrini, “La Calcolatrice Elettronica

dell’a.a. 1961/62,” internal document, September 14, 1961, AUniPi [40]. 96. “La calcolatrice elettronica e il Centro Studi

Pisana, una storia che sembra una leggenda,” in

 di Pisa,” University press release, dell’Universita

Dadda 2006, pp. 83–96, and O.G. Mancino and

November 10, 1961, AUniPi [41].

R. Sprugnoli, CEP La Calcolatrice Elettronica Pisana  Edizioni – Scenario, storia, realizzazione, eredita. Plus, 2011. 84. See Caracciolo et al. 1956 [19]. 85. B. Lojek, History of semiconductor engineering. New York, NY, USA: Springer, 2007.

97. Also thanks to Comunicato ANSA no. 86, November 9, 1961, a press release distributed by the National Press Syndicate [42]. 98. Agreement between University of Pisa and National Research Concil, July 31, 1962, AUniPi [43].

86. See, e.g., the recollection in Fabri 2013.

99. Montani et al. 2013.

87. Still in the CIU meeting of Apr. 1960, it was stated that

 svolta nel 1966,” 100. G. Capriz, “CSCE Pisa, Attivita

“while the Government finances the other universities

La Ricerca Scientifica, vol. 36, no. 2, pp. 87–94,

to buy electronic computers, Pisa is the only one that has built a large electronic computer without asking

1968. 101. The sale of Olivetti electronics division to General

anything to the State.” This refers to universities such as

Electric is one of the key event in the history of the

Turin, Padua, and Naples, and it was meant to

Italian computer industry, and the literature about it

encourage the attempts by Conversi at the Ministry and

is large: for an introduction we refer to Gemelli

at the CNR. See Minutes of the CIU Executive

2014 and for an analysis of the economic issues at

Committee, meeting of April 8, 1960, p. 58, AUniPi [36].

stake to the recent G. Monreale, Mediobanca e il

88. See Document 1958 [29]. 89. However, also this solution was late: see Minutes of the CSCE Steering Committee meeting, Feb. 17, 1961, AUniPi [37]. 90. A telegram of Christmas greetings by the CSCE researchers to Faedo (by then the new Rector), stating that the CEP is “working in the majority of its

salvataggio Olivetti. Mediobanca, 2019. 102. See Montangero 2013. 103. G.A. Cignoni, F. Gadducci, and S. Paci, “A virtual experience on the very first Italian computer,” ACM J. Comput. Cultural Heritage, vol. 7, no. 4, pp. 21:1–21:23, 2014. 104. G.A. Cignoni and F. Gadducci, “Rediscovering

instructions,” led some authors to anticipate at Dec.

the very first Italian digital computer,” in

1960 the completion of the machine: see Paladini 2009.

Proc. 3rd IEEE History Electro-technol. Conf.,

Season greetings telegram to A. Faedo, December 23,

2012, pp. 1–6.

1960, AUniPi [38]: see additional details in Cignoni, Gadducci, Ronco 2013.

April-June 2020

19

Article

Fake But True: Model Maker Roberto Guatelli, Science Museums and Replicated Artifacts of Computing History
nin Silvio He AICA, Associazione Italiana per l’Informatica ed il Calcolo Automatico

Simona Casonato Museo Nazionale Scienza e Tecnologia Leonardo da Vinci

Abstract—Science museums have a long tradition in resorting to reproductions, both to preserve original artifacts and to make exhibitions attractive by presenting the material culture of science and technology. The art of making copies of precious objects, sometimes for fakery or forgery, is as old as collecting, but since the 19th century it has become both commonplace and profitable. In this article, the authors reconstruct the life and works of a 20th century model maker, Roberto Guatelli, whose reproductions of technological artifacts of the past, particularly calculators, can be found in several museums.

& THE PURPOSE OF museums, their raison d’e^tre, is vast and complex, particularly in this age of intense debate about their nature, purpose, social, and cultural role.1 One of the oldest and most debated topics is the diverse role of

Digital Object Identifier 10.1109/MAHC.2020.2990452 Date of publication 27 April 2020; date of current version 29 May 2020.

20

1058-6180 ß 2020 IEEE

artifact collections in these institutions. There is a large body of literature on the significance of objects in museums in general,2 but science museums have their own approach. Objects that tell stories about the past of science and technology have often been considered more as “icons” and “rhetorical devices”3 rather than “significant evidence.”4 Of course, in science museums as well, historical objects are not only entities endowed with an aesthetic value or aimed

Published by the IEEE Computer Society

IEEE Annals of the History of Computing

to impress and amaze public, but they can be seen as irreplaceable documents,5 in the same way as books, manuscripts or recorded oral histories. However, in their traditional approach, science museums generally tend to display technological and industrial artifacts more to educate about contemporary science and technology, and less to illustrate historical research. Within this logic, seeing a copy (a model, an accurate replica, a physical or virtual reproduction) or seeing an authentic artifact are often construed as comparable experiences. Science museums, mostly established between the second half of the 19th century and the first half of the 20th, developed in close connections with other exhibition contexts, such as World Fairs. These fleeting events had several traits in common with museums aims (education “of the masses”), narratives (imagination of promising futures, rhetorical use of the past), and languages. Popular visual culture and the privileging of senses (sight, touch, motion, interaction) have always been important components of the culture of scientific exhibitions, even at their remote origins.6,7 The practice of replicating object is deeply connected to science museums. It is true that, on the one hand, in the permanent museums, technological artifacts gathered for temporary exhibitions face the challenge of their conservation. Resorting to physical “copies” of the original artifact, faithful enough to the smallest details, can be a solution when a historical relic is particularly perishable, and it may be difficult to combine the needs of conservation with those of exhibition, as in the world of art.8 At the same time, the World Fairs’ tradition of creating copies of unavailable original objects, has been continued by science museums, which are generally less interested in the authenticity of iconic artifacts than in their presence, even as copies, in their collections. Several cases of replicating historical objects feature prominently in the history of computing, e.g., the Antikythera mechanism, Babbage’s Difference Engine, Pascal’s dial adder, Colossus, Turing’s Bombe, Manchester Baby, just to name a few.9 There is no univocally established terminology in the literature to characterize these operations, so we propose a practical one, debatable as it is any taxonomy: “replica” and “reproduction” can be considered synonyms, in reference to a copy

April-June 2020

as close as possible to an extant artifact, regardless its functionality. In contrast, “reconstruction” refers to an artifact which is a copy of an object built in the past, but lost or severely damaged, of which we still have fragments, photographs, or blueprints, complete and precise enough to attempt remaking the object. Finally, there are the ‘interpretative models’,10 realizations of machines that were conceived and designed in the past, but never built, of which we are left with literary descriptions, rough drawings or sketches only. These should be thoroughly studied to know if the machines they refer to could have been really built and could have worked. In such cases, drawings and descriptions are subjected to a hermeneutic process, considering the frame of knowledge and technology of the inventor’s time. Well-known examples are the interpretative models made after the many sketches of Leonardo da Vinci, or the Difference Engine No. 2 by Charles Babbage.11,12 In many of these cases, we notice that although reconstruction practices have been originated by a culture of exhibition, they have also another important purpose, in that they can also be paired to research in the field of “experimental archaeology.” This term refers to the study of all aspects concerning the artifact in its historical frame, such as producibility, economics, ergonomics, practical usability etc. If this is the case, the faithful reproduction of the object’s aesthetic is not enough, but there is the need to use the same materials and to use the production techniques available at the time of the original. The physics of the original artifacts should be respected, not only its geometry, so scale models are hardly reliable. Of course, the difference between “reconstruction” and “interpretative model” may be fuzzy, depending on the meaning we give to the words “documentation complete and precise enough.” In the case of Colossus, an interpretative study was necessary, even if a dozen machines were built and used, due to the shortage of available documents. Models of the Antikythera mechanism were also built, but they are still speculative, after decades of intensive studies, because scholars can only rely on a few degraded fragments and on a few lines of text. 13 With this brief and partial list of considerations we wish to point out that the art and

21

Article

science of replicating ancient technologies is a multiform field of enquiry. It encompasses a range of positions and attitudes, from potential fakery to rigorous scientific research. The proliferation of small museums and temporary exhibitions, as well as the massification of collecting, and even souvenirs sold in museum shops, created an attractive business for willing craftsmen and entrepreneurs. With respect to the technological and scientific domain, the history of exhibit designers and model makers, a topic seldom explored, can provide interesting results.14 This article reconstructs the life and work of one of them, who began operating in the 1940s producing models of Leonardo da Vinci’s mechanisms and went on for more than four decades building reproductions of historical computing machines.

ROBERTO AMBROGIO GUATELLI (1909–1993) Little has been published so far on Roberto Guatelli, a skilled craftsman who built many replicas and models of ancient machines, now dispersed in science and technology museums in European and American countries. For the most part, the published sources about him are very brief articles by scholars and museum curators, such as Jim Strickland,15 Erez Kaplan,16 David Pantalony,17 and C. Boyer.18 Only the last author could interview Guatelli, who died in 1993, the others could only draw on the memories of Guatelli’s stepson and collaborator Joseph F. Mirabella Jr., who is still alive. The main interest of these scholars was, however, limited to some replicas and reconstructions made by Guatelli in the years 1950–1980s, in particular the fancy interpretation of a presumed ‘Leonardo’s calculator’ that Guatelli built in 1968.15,16 Roberto A. Guatelli was born in Binago, a small town 50 km North of Milan, on December 4, 1909,19 son of Andrea Guatelli, a furniture manufacturer. According to a 1949 press release,20 at a young age Guatelli became passionate about the machines designed by the Italian Renaissance master Leonardo da Vinci. Guatelli’s father had purchased an expensive reproduction the Leonardo’s Codex Atlanticus, the original of which is kept at the Biblioteca Ambrosiana in Milan. In

22

the volume the young Roberto discovered sketches of mechanisms and machines, for civil and military use, that fascinated him. According to the press release, Guatelli “graduated in Civil Engineering and Electrical Construction at the University of Milan” in 1931. In the 1930s he took care of his father’s business and travelled across Europe, before taking a job at Azienda Elettrica Municipale (Milan Power Company) where he worked on the design of hydroelectric plants in the Alps. His love for Leonardo was gratified in the late 1930s, when “[T]he City Council of Milan requested Guatelli to organize a commission to prepare blueprints from Leonardo’s drawings in order to construct a series of models illustrating the scientific achievements of Leonardo da Vinci [. . .] Roberto had been supervising the Commission which had been preparing the blueprints for the 124 scale models which were to comprise the Exhibit.”20 Other sources go so far as to say that “Benito Mussolini commissioned Roberto Guatelli, an Italian modeler/engineer, to create scale models from the sketches” and to call him “Dr. Guatelli, a world expert of Leonardo.”18,21 Those statements, which were repeated several times in the American press, are rather exaggerated, as we shall see. For the moment it suffices to say that, almost certainly, Guatelli did not graduate or enrolled at university,22 and that he was definitely not commissioned by Mussolini to organize the exhibition, nor did he design any model for it. In 1937, the City Council of Milan, under the aegis of the Fascist Government, organized a great event to celebrate Italian inventiveness with the aim to carry out nationalist propaganda at home and abroad.23,24 The figure of the great Leonardo, both artistic and scientific, fitted perfect for this occasion. A scientific committee and an executive committee were organized, the last headed by the superintendent of the Milanese Civic Museums, Giorgio Nicodemi. The exhibition was set up at Palazzo dell’Arte in Milan and was open from May 9 to October 22, 1939. It included almost 200 models of Leonardo’s machines, many in a 1:1 scale, some of which were functional and could be manipulated by visitors. The models were also painted in bright colours to facilitate visitors’ understanding of how they worked.25–27 The exhibition attracted a

IEEE Annals of the History of Computing

great interest also from abroad. Some cultural institutions thought buying the entire collection, which would have allowed the City of Milan to balance the high costs of putting on the exhibition.26 The Italian Ministry of Popular Culture, the body responsible for fascist propaganda, decided instead to send the exhibit to the USA, on occasion of the 1939–1940 World Fair, to be held in New York. The transfer of the entire collection to New York had to be organized and here the name of Roberto Guatelli appears for the first time, as he was commissioned to take care of the models’ shipment, their installation, and maintenance, with the help of Armando Pistoso.27 The materials were divided and sent on two ships, the luxurious Conte di Savoia, on which Nicodemi, Guatelli and Pistoso also sailed,28 and the cargo Barbarigo.29 The latter reached New York first, on May 14, 1940, while the former arrived nine days later. The exhibition was hosted at the Museum of Science and Industry in the Rockefeller Center, New York, and was held from July 1940 to January 1941, achieving great public and media acclaim. Rich patrons and cultural institutions, including the Rockefeller Foundation, wanted to buy the models or at least to rent them for other exhibitions in the USA.29 Guatelli’s work was undoubtedly crucial for the success of the event. The young artisan, who did not know a word of English, was helped by Florence Nonna, an Italo–American living in New York City, whom he married after the war. At the end of the exhibition, Guatelli wanted to stay in the USA and asked a lawyer named Joseph Mirabella Sr. for legal help to get temporary residence.30 Italy was already at war on the side of Germany, but not against the United States, which was still neutral. The Italian government decided instead to send the exhibit to Japan, to ensure good relations between the two Axis countries, and to order Guatelli and Pistoso to go with it. A few weeks before Pearl Harbor, in September 1941, the models were sent on two Japanese steamships, the Awajisan Maru, which left from New York, and the Tatsuta Maru, which sailed from San Francisco. Guatelli and Pistoso travelled on the former. The exhibition was held in Tokyo from July to September 1942, with the incongruous title of “The Asian [sic] Renaissance.

April-June 2020

Exhibition by Leonardo da Vinci,” and was intended as an instrument of Japanese war propaganda.31 After the Tokyo exhibition, the collection should have been sent back to Italy, but by then the USA and Japan were at war and the Soviet Union was also at war with Germany. It was therefore impossible to ship the models to Manchuria and use the Trans-Siberian Railway to reach Italy, but Pacific Ocean and Indian Ocean were also dangerous routes in wartime. On the fate of the collection sources are discordant, some authors say it was burned by Allied bombers while in storage at an embassy in Tokyo, waiting to be shipped,15,18,20,32 others claim that it was lost in the wreckage of the ship carrying it.25,33,34 From a recent testimony it seems that the collection was perhaps loaded on a Japanese submarine, which was sunk in the Indian Ocean.31,35 Be that as it may, the entire collection of models was lost. According to the press release mentioned above,20 Guatelli learned Japanese during his stay in Japan and wrote a book in that language about Leonardo with Vinzo Comito. It seems unlikely that, in less than two years, they mastered Japanese well enough to write a book, but the volume does indeed exist and carries their names as authors.31,36 It is also reported that Guatelli was knighted by Emperor Hirohito,20 but there is no further evidence of this fact.37 When Italy surrendered to the Allies in September 1943, the Japanese considered all Italian residents as traitors and interned them in concentration camps, and this was also the fate of Guatelli and Comito.38 However, it seems that both were not treated too harshly and enjoyed a certain amount of freedom, thanks perhaps to the imperial honour.20 Allied occupation forces freed the Italian prisoners in September 1945.39 Guatelli, who had been missing from Italy for five years, probably returned home, but he went back to the US again in 1948, where he was reunited with Florence and went to live in her parents’ house in New York City. During the War, Florence, believing him missing, had married the lawyer Joseph F. Mirabella Sr. and had a son, Joseph F. Mirabella Jr.40 Around this time, Guatelli heard that Wilhelm R. Valentiner of the Los Angeles County Museum was looking for the 1939 collection of Leonardian

23

Article

models. Valentiner wanted to display them in a new exhibition, not knowing that they were lost. Guatelli saw an attractive opportunity for himself and proposed to Valentiner to build a new set of models based on the 1939 blueprints. An agreement among Guatelli, the Los Angeles Museum and Panold Masters Inc. Ltd. (a company operating in the art market), made it possible to reconstruct in record time 65 new models of Leonardo’s machines.41 Panold Masters financed the project, got the ownership of the models and paid Guatelli as consultant.20,42,43 The new exhibition ran from June 3 to July 27, 1949. It was a success, receiving more than 100,000 visitors. Panold Masters’ publicity encouraged other museums to request the collection, which went on to be shown at the Dallas Museum of Fine Art, the Denver Civic Center, the Ford Institute Museum in Dearborn (MI), and the B€ uhl Planetarium in Pittsburgh.43 It was from these exhibitions that Guatelli’s name rose to fame, at least in the US. Public interest ran out after the exhibition in Pittsburgh. In 1950 van Riemsdyk, a manager of Panold Masters, sent a letter to the president of the Ford Institute, complaining that loan revenues did not compensate for the costs incurred and that Guatelli “[. . .] whose livelihood depends on the continuation of the present exhibition” was concerned.43,44 Fortunately, Guatelli had been introduced to the president of IBM Thomas Watson.15–18,45 Watson, a passionate collector of art and antiques, decided to buy the entire collection from Panold Masters and hired Guatelli to build more models. In 1951, Guatelli moved to a workshop in the IBM World Headquarters, 590 Madison Avenue, NY.17 IBM organized a new exhibition entitled Leonardo da Vinci: An Exhibition of his Scientific Achievements, Collection of the Fine Art Department, International Business Machines Corporation, which was shown for the first time at the IBM Country Club in Endicott (NY) in July 1951 and at the IBM offices in Poughkeepsie (NY) in August.29,34 In the following years, a travelling exhibition was organized and was shown in many American cities between 1952 and 1954, beginning with the Metropolitan Museum in New York.29 The models were also sent to Paris (France) in 1951–1952, accompanied by Mrs. and Mr. Guatelli, and displayed

24


at the Palais de la Decouverte, on the occasion of the 500th anniversary of Leonardo’s birth.46 In 1954, Guatelli obtained American citizenship47 and, seven years later, he left IBM to open his own company, the Roberto A. Guatelli & Associates Inc., based in Lafayette Street, New York City. In 1964, his stepson Joseph Mirabella Jr. joined him in managing the company.40 Guatelli and Mirabella went on to build models and reconstructions for IBM, in order to enrich the Fine Art Department, but they also worked for other companies and perhaps for museums. Roberto Guatelli died in 1993, Mirabella continued his stepfather’s activity until 2005.40

HERITAGE OF GUATELLI AND MIRABELLA. After being hired by IBM, Guatelli began building several further replicas, reconstructions and interpretative models, at first only for IBM, afterwards also for other costumers. He certainly went on with Leonardo’s models, in fact, their number increased to 75 in 1951, when they were exhibited in Philadelphia.48 In an email to Valery Monnier, Mirabella asserted that at least two inventories of the replicas built in their workshop were made, but both seem no longer available.49 According to Boyer,18 Guatelli and Mirabella had built between 600 and 2000 replicas and models until 1984. The range is large, but even taking for granted the lowest figure, it is only natural to wonder what happened to them all. IBM’s Fine Arts Department disappeared for more than two decades. An article by Chin-Tao Wu alleges that “IBM not only closed its Gallery of Science and Arts in its New York headquarters in August 1994, but, most significantly, sold off a substantial part of its art collection.”50 However, we have found conflicting statements to this regard. As mentioned in the book A Calculation Chronicle. 300 Years of Counting and Reckoning (IBM, 1990), the collection also included two precious pieces: one original Pascal’s dial-adder of 1647 and the wonderful Thomas de Colmar’s piano-arithmometer of 1855. Both seem to have disappeared; this may have also been the fate of the models and replicas made by Guatelli. Strangely enough, the website of the IBM

IEEE Annals of the History of Computing

Corporate Archives collection in Poughkeepsie is still online, showing 127 artifacts, ranging from ancient abaci to modern electronic computers.51 Regretfully, the images are low quality, with a few lines of description, no attribution, and it is not specified whether the objects are originals or replicas. In particular, the collection owns six models of Leonardo’s machines, which are perhaps relics from Guatelli’s 1949 set. Some of the Leonardo’s models have previously been donated to museums, such as the Museo Leonardo in Vinci (Leonardo’s hometown
Castle near Florence, Italy)52 and the Clos-Luce Museum (near Amboise, in France, da Vinci’s home during the last final years of his life).53 It is not known if they were part of Guatelli’s original set, the ones from IBM’s travelling exhibition, or if they were built afterwards. In 1951, the IBM Italian subsidiary offered 35 models to the newly founded Museo della Scienza e della Tecnica in Milan. The construction would have been commissioned to Guatelli’s brother, who lived in Milan. However, the proposal was not accepted because Guatelli’s interpretations were considered “unscientific” by Italian scholars, moreover Guatelli did not have a good reputation in Italy: when the City Council of Milan, owner of the 1939 drawings, heard about the exhibition in Los Angeles, they considered suing Guatelli for plagiarism.54 The design and construction of the new Milanese collection was commissioned instead to workshops run by the Italian Armed Forces.25 Recently, Joseph Mirabella has donated other 39 models to the Long Island Science Center (Riverhead, NY).55 Guatelli is perhaps best known in our times for his hypothetical model of a calculator invented by Leonardo da Vinci. The story was recounted
nin.56 In by Kaplan,16 and more recently by He 1967, Guatelli saw a reproduction of Leonardo’s Codex Madrid I exhibited in Boston. His attention was caught by a sketch of a gear train, which he interpreted as an adding machine, of the same type as Pascal’s dial adder. He made a model of it, which was exhibited at the IBM Science and Art Gallery. Experts soon criticized the interpretation, and “[. . .] an academic trial was then held at the University of Massachusetts [. . .]. Among others, Prof. I. Bernard Cohen, consultant for the IBM collection, and Dr. Bern Dibner, a leading

April-June 2020

Leonardo scholar, were present.” The conclusion was that “Such a machine could not be built due to the enormous amount of friction that would result. [. . .] The vote was a tie, but IBM decided to remove the controversial replica from its display.”16 There are other reasons to reject Guatelli’s interpretation. In brief: 1) the original Leonardo’s sketch does not show a device to stop the wheels in ten discrete positions, 2) no way to enter addends is shown, and 3) on the first and last axis two weights are hanging, which would have no function in a calculator. In fact, Guatelli had to modify the Leonardo sketch quite a bit to overcome those inconsistencies. But the main reason is Leonardo’s own hand-written notes under the sketch, where he explains how gear trains can be used to increase force.56 Guatelli cannot be blamed, because the first printed transcription of Codex Madrid was published six years later and it was in XV century Italian, a language that even Italian speakers can hardly understand. From the 1960s to the 1980s, IBM commissioned Roberto Guatelli & Associates to reproduce some of the most famous early mechanical and electromechanical calculators, among them:  











Blaise Pascal’s Pascaline (1645);57 Gottfried Wilhelm Leibniz’s stepped reckoner (ca. 1694);58 Charles Babbage’s Difference Engine N. 1 (1832);59 Herman Hollerith’s punched card tabulator and sorter (1890);60 Babbage’s trial model of the Analytical Engine (1871);61 Georg and Edvard Scheutz’s Difference Engine N. 2 (1859);62 Giovanni Poleni’s Machina Aritmetica (1709).63

Herman Bruderer, who recently published two short accounts about Guatelli’s replicas, suggests that Guatelli and Mirabella may also have built copies of other calculating machines, such as the Millionaire, the Adix, and Webb’s adding machine.64 Moreover, there is evidence that, after 1961, Guatelli built models for other customers: at least one artifact for the Digital Equipment Corporation, a full-scale replica of Hollerith’s 1890 tabulator,65 and an interpretive model of Antonio Meucci’s telephone blueprints

25

Article

Figure 1. Gottfried Leibniz’s stepped reckoner (1690–1710), replica by R. Guatelli (Picture: Museo Nazionale Scienza e Tecnologia Leonardo da Vinci, Milano / A. Nassiri).

for Anthony De Nonno, a filmmaker and passionate scholar of Meucci’s life.66 Another reconstruction, mentioned by Boyer,18 is the Selective Sequence Electronic Calculator (SSEC), a huge hybrid computer built by IBM in 1944.67 This seems rather improbable, the SSEC was too

Figure 2. Charles Babbage’s Difference Engine N. 1 (1832), replica by R. Guatelli (Picture: Museo Nazionale Scienza e Tecnologia Leonardo da Vinci, Milano / A. Nassiri).

26

complicated and its reconstruction would have cost too much, but it is possible that Guatelli restored an extant part of it. Of all those reproductions, no more than a score are traceable today and can be found in the Traub–McCorduck Collection at the Carnegie Mellon University,64,68 the Computer History Museum in Sunnyvale (CA), 64,69 the Canada Science and Technology Museum, Ottawa (Canada) 64,70 and the Museo Nazionale Scienza e Tecnologia Leonardo da Vinci, Milan (Italy) 64,71 (see Figures 1–3). The Arithmeum of Bonn (Germany) probably has the richest collection, owning seven replicas by Guatelli, two of which are demonstration pieces of calculator mechanisms.72 It seems that IBM management also used to give models and replicas to subsidiary headquarters worldwide, which in turn donated them to museums. In fact, one replica of Babbage’s Differential Engine N. 1 is still at IBM Italia premises in Segrate (Milan, Italy).73 It cannot be excluded that models could have also been given to retiring IBM managers, such as the replica of Babbage’s Difference Engine N. 1, which was given as a present to T. Vincent Learson––well-known creator of IBM System/360, by his colleagues. This replica is now at the Computer History Museum of Sunnyvale, CA. It is difficult to reconstruct the story of every single replica, as data are scant, incomplete, and often contradictory. It is also worth noting that, in general, the attribution cannot be taken for sure, as Guatelli and Mirabella did not brand their products with trademarks or other signs, as far as we know. A reasonable assumption seems to be that all models and replicas donated

IEEE Annals of the History of Computing

Figure 3. Blaise Pascal’s dial adder (Pascaline, 1645), replica by R. Guatelli (Picture: Museo Nazionale Scienza e Tecnologia Leonardo da Vinci, Milano / A. Nassiri).

by IBM should be attributed to Guatelli or Mirabella, because it seems improbable that the corporation had more than one supplier for these sorts of objects. There is, however, at least one exception: the replica of Babbage’s Difference Engine N. 1 owned by the National Museum of American History in Washington D.C., which was donated by IBM in 1963.74 In the Museum catalogue the object is attributed to Daniel I. Hadley & Associates, a model maker specialized in architectural models. In fact, the reproduction is in scale 1/1, whereas other Difference Engine replicas by Guatelli were made in smaller scale (0,56/1). One cannot even assume that all replicas attributed to Guatelli were given by IBM to the current owner, as the artifacts at the Arithmeum were sold directly by Guatelli to the museum.75

CONCLUSION The reproductions of Guatelli and Mirabella vary greatly in terms of historical accuracy, completeness and operability. At the higher end, there are the replicas of Pascal’s dial-adder, built in full scale, working, and accurate enough to be confused with the original by nonexperts, at least at a first glance. At the other end, we might put the copy of Leibniz’s stepped-drum reckoner, made in reduced scale, missing all the internal mechanism, in practice just a mock-up, utterly implausible. Between these extremes, we can place all the others, at least the known ones. To this variability certainly contributed time and cost constraints imposed by customers, whose purpose was the promotion of the corporation’s image, rather than the historical fidelity.

April-June 2020

Nevertheless, Roberto Guatelli and his stepson Joseph Mirabella were certainly skilled craftsmen who understood the potential and promising market of reconstructions, models, and replicas, which afterwards has further expanded into an interesting business, not only as objects to be displayed in museums, but also as souvenirs to be sold to visitors and collectors. At times their interpretations of ancient drawings and projects may have been fanciful or philologically incorrect, and so their models can be criticized as is any interpretation of the past. British novelist Lesley P. Hartley said: “The past is like a foreign country: they do things differently there.”76 From a scholarly point of view, of course, the making of replicas and reconstruction should follow some good practices and scientific criteria. In order to provide an effective experience, they should not be roughly and coarsely built, taking care of the aesthetic value only, but should respect the original sizes, to achieve the “feeling” of volume and mass, which can be perceived even without touching the artifact. This rule is difficult to follow when the historical artifact is too big or the cost would be excessive, as it happens for instance in the case of interpretative models built by experimental archaeologists. Even if this is not the case of machines and instruments, such as clocks, calculators, measuring instruments, experimental setups, etc., in which the size is not the major obstacle, replicating delicate mechanisms, especially with the aim of creating working machines could be equally difficult, time consuming and expensive.

27

Article

Beside practical constraints, the variability in Guatelli and Mirabella’s works should also be read in terms of cultural history. This is an interesting account of the interaction between private and public actors in disseminating the history of science and technology thanks to the medium of the exhibition. In the circulation and market of replicated artifacts, we can see an ongoing dialogue between the “fair” and the “museum,” the industrial promotion and the scientific education. Indeed, in 1952 the Italian scholars that rejected Guatelli’s Leonardian models argued that they could be “interesting for IBM advertising, but they are not accurate documents fit to the Leonardo exhibition,” acknowledging the habit of corporate communication throughout artifacts.54 In this sense, modern replicas, models, reproduction, and reconstructions have a different historical value from original artifacts. Nevertheless, they should be studied with the same attention, documenting their origin, their “biography” and their conservation condition. When acquired or produced by permanent institutions, such as science museums or archives, they could achieve the dignity of originals. Of course, these copies do not only replace the original objects that inspired their (re)production, but also shift their “aura,” to put it in Walter Benjamin’s terms, the feeling of uniqueness and direct connection with the past that emanates from authentic works of art and that is lost in mechanically reproduced copies.77 Clearly, the art of artisanal replication explored here is different. Guatelli’s history shows that reconsidering the aura of reproductions could be an interesting exercise, consistent with the ongoing process of reinterpretation the role of science museums and of public history of science in the contemporary age.78

born, MI), Los Angeles County Museum (Los Angeles, CA), Balch Art Research Library (Los Angeles, CA), Dallas Museum of Art (Dallas, TX), Denver Art Museum (Denver, CO), Museo Nazionale Scienza e Tecnologia Leonardo da Vinci (Milan, Italy), Museo Leonardiano (Vinci, Italy),
- Parc Leonardo da Vinci Chateau du Clos Luce (Amboise, France), Long Island Science Center (Long Island, NY). We are indebted to Gerardo Con Diaz for his help with the style and English revision.

& ENDNOTES/REFERENCES 1. See, for instance, the International Council of Museums (ICOM)’s debate about the new definition of ‘museum’. Accessed: Mar. 5, 2020. [Online]. Available: https://icom.museum/en/standardsguidelines/museum-definition/. 2. A. Desvallé es and F. Mairesse, Key Concepts of Museology. Paris, France: Armand Colin, ICOM International Councils of Museums, & Museé Royal de Mariemont. 2010; S. H. Dudley (ed), Museum Objects, London-New York, NY, USA: Routledge, 2012. 3. R. Bud, “Museums theme–Adventures in Museology: Category building over a century, and the context for experiments in reinvigorating the Science Museum at the turn of the twenty-first century,” Sci. Museum Group J., no. 08, 2017, doi: 10.15180/170809. 4. From the Artefacts Consortium’s website. Born in 1993, this group of scholars and curators study the cultural and historical significance of objects in science museums. Accessed: Mar. 5, 2020. [Online]. Available: http://www.artefactsconsortium.org/ AboutUs/MainAboutUsF.html. 5. Original, Copy, Fake, On the significance of the object in History and Archaeology Museums, ICMAH – ICOM International Committee for Museums and Collections of Archaeology and History, 22nd General

ACKNOWLEDGMENTS For the encouragement given to the writing of this article and for the valuable comments to the manuscript, we warmly thank Michael R Williams, Dag Spicer, and Doron Swade. For help in finding precious information we are grateful to Andrea Bernardoni, Herbert Bruderer, Davide Colombo, Claudio Giorgione, Bernhard Korte, Joseph Mirabella, Valery Monnier, David Pantalony, Carlo Randone, Luca Reduzzi, Yoshimi Takuwa, Ed Thelen, The Henry Ford Org (Dear-

28

Conference, Shangai, China, Nov. 2010. 6. E. Canadelli, M. Beretta, L. Ronzon (eds), Behind the Exhibit. Displaying Science and Technology at World’s Fairs and Museums in the Twentieth Century. Washington DC, USA: Smithsonian Inst. Scholarly Press, 2019. 7. S. J. M. M. Alberti, “Objects and the Museum,” Isis, vol. 96, no. 4, pp. 559–571, 2005. A. Griffiths, Shivers Down Your Spine. Cinema, Museums and the Immersive View. New York, NY, USA and Chichester, U.K.: Columbia Univ. Press, 2013.

IEEE Annals of the History of Computing

8. A. Riegl, Der Moderne Denkmalkultus, Seine Wesen

19. Some authors state that Guatelli was born in Milan

Und Seine Entstehung, Vienna, Austria, 1903, quoted

in 1904, but birth date and place can be found on

in S. Barassi, “The modern cult of replicas: A Rieglian

three documents: U.S. Naturalization Records

analysis of values in replication,” TATE’s Online Res. J.

Indexes of 1954, New York State, Passenger and

Accessed: Jan. 23, 2020. [Online]. Available: https://

Crew List, of May 23, 1940, and the New York

www.tate.org.uk/research/publications/tate-papers/

State, Passenger and Crew Lists of Flight TWA 777

08/the-modern-cult-of-replicas-a-rieglian-analysis-ofvalues-in-replication. 9. D. Swade, “Reconstructions as experimental history.

from Paris, 1952. Source: Ancestry.com. 20. Edison Institute, News Release, How the Leonardo da Vinci Exhibit Became a Reality, 3 January 1950,

Historic computing machines,” in K. Staubermann

Archives of The Henry Ford Institute, Dearborn (MI),

(ed.), Reconstructions. Edinburgh, Scotland: National

EI 235 Box 1 Folder 11 (Courtesy of The Henry Ford,

Museum of Scotland, 2011, pp. 103–126.

Sam Rod).

10. ‘Reconstruction’ (from the Latin re-construire – to make

21. For instance: G. Mac Cullum, “Interesting people

again) should not, in principle, be used for objects

put science in the background,” Dallas Morning

never built before.

News, Oct. 15, 1949; “Updating da vinci,” The

11. In some cases, we could speak of “interpretive

Rotarian, Dec. 1952, p. 14; “Doctor’s boyhood

reconstructions”, as for instance the Astrarium by

hobby of model building pays off,” Sunday Gazette.

Giovanni Dondi. The original and working instrument

Charleston, West Virginia, Feb. 7, 1960; K.

was created in the 14th century and lost afterwards. It

Mckenna, “Art review: Da vinci: Father of invention,”

was reconstructed it in 1961-63, thanks to the

Los Angeles Times, Apr. 4, 1989; F. C. Moon, The

interpretation of precise instructions left by Dondi

Machines of Leonardo da Vinci and Franz

himself. See C. Zanetti and J. Torriani, A Reinassance

Reuleaux, Springer, 2007, p. 201; B. Atalay, Math

Genius. Cremona, Museo del Violino, 2016, Art. no. 240.

and the Mona Lisa. The Art and Science of

12. D. Swade, The Difference Engine: Charles Babbage

Leonardo da Vinci, Smithsonian Books, 2007,

And the Quest to Build the First Computer. Baltimore,

pp. 203–204; “Leonardo airborne and flights of

MD, USA: Penguin, 2000.

fancy,” Madame Pickwick Art Blog, Aug. 25, 2009,

13. A. Jones, A Portable Cosmos. London, U.K.:Oxford Univ. Press, 2017. 14. TATE Gallery, Replica. A copy of a work of art that is virtually indistinguishable from the original. Accessed:

Accessed: Feb. 2, 2020. [Online]. Available: http:// www.madamepickwickartblog.com/2009/08/ leonardo-airborne-and-flights-of-fancy. 22. The

only

university

in

Milan

which

issued

Feb. 21, 2020. [Online]. Available: https://www.tate.

engineering degrees was, and still is, the

org.uk/art/art-terms/r/replica.

Politecnico, but the name Guatelli does not appear

15. J. Strickland, “Who was that Guy?”, Comput. History

in the lists of ex-alumni (Historical Archives of the

Museum. Volunteer Inf. Exchange, vol. 2, no. 3,

Politecnico, Milan, personal communication,

pp. 2–3, Feb. 15, 2012.

Jan. 2020). In fact, in an Italian newspaper of 1940

16. E. Kaplan, “Leonardo’s calculator,” ETCetera –

Guatelli was called ‘perito’, a technical high school

Mag. Early Typewriter Collectors Assoc., vol. 37,

diploma akin to Grade 12 in the US (“La

pp. 6–7, Dec. 1996; E. Kaplan, “The controversial

Leonardesca a New York”, Corriere della Sera,

replica of Leonardo da Vinci’s adding machine,” IEEE Ann. History Comput., vol. 19, no. 2, pp. 62–63, Apr.–Jun. 1997. 17. D. Pantalony, “Collectors, displays and replicas in

April 26, 1940, p. 4). 23. M.

Beretta,

E.

Canadelli,

and

C.

Giorgione,

Leonardo 39. La costruzione di un mito, Editrice Bibliografica, 2019. Manufacture and preparation

context: What we can learn from provenance research

were commissioned to the Ministero

in science museums,” in J. Buchwald, L. Stewart (eds.),

dell’Aeronautica, Ministero della Guerra, R.A.C.I.

The Romance of Science: Essays in Honour of Trevor H.

(Reale Automobile Club d’Italia), R.U.N.A (Reale

Levere, Berlin, Germany: Springer, 2017, pp. 255–275.

Unione Nazionale Aeronautica), Salmoiraghi, and

18. C. Boyer, “Leonardo’s man in SoHo,” THINK, pp. 41–43,

others (Archivio Storico Civico, Milan, Archivietto

Nov./Dec. 1984.

April-June 2020

Rivolta, Cart. 41).

29

Article

24. In those years, Italy was sanctioned by the League of Nations for the assault on Ethiopia. Though not particularly severe and not respected by all countries, the sanctions were exploited by Fascism to stir up patriotism and reinforce public support. 25. C. Giorgione, “The birth of a collection in Milan: From

works at Endicott and Poughkeepsie,” IBM World Trade News, vol. 3, no. 8, Aug. 1951. 35. It seems unlikely that tens of crates could be loaded in a claustrophobic submarine, but, since 1943 the Japanese Navy converted some war

the Leonardo exhibition of 1939 to the opening of the

submarines into cargos that could load up to 100 –

national museum of science and technology in 1953,”

1000 tons, to send precious minerals to Germany in

Sci. Museum Group J., Autumn 2015, no. 4, 2015.

exchange for technology (The United States

Accessed: Feb. 21, 2020. [Online]. Available: http://

Strategic Bombing Survey, The War Against

dx.doi.org/10.15180/150404/002

Japanese Transportation (1941–1945),

26. R. Cara, “La mostra di Leonardo da Vinci a Milano tra arte, scienza e politica (1939),” in M. Toffanello,

Transportation Division, May 1947, p. 63). 36. V. Comito and R. Guatelli, Leonardo da Vinci, Genio

All’origine delle grandi mostre in Italia, Il Rio Arte,

Universale. Tokyo, Japan, 1942. Vinzo Comito was

2017, pp. 137–161.

an Italian journalist who worked in the USA in the

27. A. Bernardoni, “Le origini della filologia macchinale

Thirties to spread fascist propaganda to Italo-

vinciana. Modelli per il volo (1929-1939),” in

Americans. In 1941 he fled to Japan, where he

M. Beretta, E. Canadelli, and C. Giorgione, Leonardo

continued that activity. (V. Comito, It was a

39. La costruzione di un mito. Milano, Italy: Editrice

revolution, Shanghai, XX Century Publishing Co.

Bibliografica, 2019, pp. 97–116.

1942; Hearings before the Select Committee to

28. New York State, Passenger and Crew List, Conte di Savoia, May 23, 1940. Source: Ancestry.com 29. D. Colombo, “La Leonardesca in trasferta a New York.

Investigate the Federal Communication Commission, House of Representatives, 78th Congress, First Session, Part 1, p. 304; Second

Un modello per le mostre delle macchine vinciane

session, Part 3, p. 2682; “FCC Explains

negli Stati Uniti,” in M. Beretta, E. Canadelli, and

Background of Activity in Foreign Language

C. Giorgione, Leonardo 39. La costruzione di un mito.

Broadcast Field.” Broadcasting, August 9, 1943,

Milano, Italy: Editrice Bibliografica, 2019, pp. 169–190.

p. 48).

Colombo names the Agostino Barbarigo, which was an

37. Y. Tukawa, Personal communication, Jan. 2020.

Italian Navy submarine. There was, however, a cargo

38. Only the few who swore allegiance to Japan and to

ship Barbarigo of the Venetian Society of Navigation.

Mussolini were saved. Yoshimi Tukawa kindly sent

Accessed: Feb. 18, 2020. [Online]. Available: https://

me the copy of a hand-written note by Guatelli and

www.wrecksite.eu/wreck.aspx?97192.

Comito, dated July 25, 1945 and sent from

30. J. Mirabella, Personal communication, Jan. 20, 2020. 31. Y. Takuwa, “Leonardo da Vinci a Tokyo nel 1942. La Leonardesca tra propaganda di guerra e Giappone

Sengoku-hara, which was the site of an internment camp. Nothing is known about the fate of Mr. Pistoso. 39. P. D’Emilia, “La guerra dimenticata tra l’Italia e il

postbellico,” in M. Beretta, E. Canadelli, and

Giappone,” BUROGU, occhi sull’impero. Accessed:

C. Giorgione, Leonardo 39. La costruzione di un mito.

Jan. 30, 2020. [Online]. Available: https://burogu00.

Milano, Italy: Editrice Bibliografica, 2019, pp. 191–206.

wordpress.com/2015/05/12/la-guerra-dimenticata-tra-

32. Boyer says it was the Italian Embassy, while Joseph

italia-e-giappone-di-pio-demilia-prima-parte/

Mirabella remembers it was the British Embassy

40. J. Mirabella, Personal communication, Jan. 2020.

(J. Mirabella, personal communication). In the wartime,

41. Other sources give different numbers, from 56 to 68.

the British Embassy was a hospital for prisoners of war

42. W. R. Valentiner and W. E. Suida, Leonardo da Vinci

and was not bombed M. M. Whiteman, Digest of

Loan Exhibition. Los Angeles County Museum, 3 June

International Law. Washington, D.C., USA: US Dept.

– 17 July 1949. Los Angeles, CA, USA: Los Angeles

State, 1970, Art. no. 465.

County Museum, 1949, Art. no. 12.

33. C. Giorgione, Leonardo da Vinci, la collezione dei

30

34. “IBM presents model display of Leonardo da Vinci

43. Letter of R. van Riemsdyk (Panold Masters) to B. Ford

modelli del museo. Milan, Italy: Edizioni Fondazione

(Edison Institute), May 18, 1950. Archives of The

del Museo Nazionale Scienza e Tecnologia Leonardo

Henry Ford Institute, EI 131 Box 18 Folder 29,

da Vinci, 2009, Art. no. 19.

(Courtesy of The Henry Ford, Sam Rod).

IEEE Annals of the History of Computing

44. The loan cost US$ 3000, inclusive of transport and the

53. C.-L. Parc and L. da Vinci, Accessed: Jan. 12, 2020.

fee for Guatelli and his assistant (Archives of The

[Online]. Available: http://www.vinci-closluce.com/en.

Henry Ford Institute, EI 131 Box 18 Folder 29). After

Catherine Cuellar, personal communication,

five exhibitions, Panold Masters’ income should have

Jan. 2020.

been about US$ 15,000 (US$ 158,000 in 2020,

54. G. Ucelli, Promemoria, 28/12/1951. Archivio Storico

Measuring Worth. Accessed: Jan. 12, 2020. [Online].

Museo della Scienza e della Tecnica di Milano,

Available: https://www.measuringworth.com/.

Archivio Museo, Corrispondenza, I serie, 1927–1964,

45. Strickland15, perhaps based on an interview with Mirabella, states that Guatelli was introduced to

b. 32. 55. C. Kaller, Long Island Science Center, Grand

Watson by Eleanor Roosevelt, wife of the late

Opening. Accessed: Jan. 20, 2020. [Online].

president, while in Dearborn. We have not found any

Available: https://www.sciencecenterli.org/latest-

evidence of this.

news/2019/1/8/long-island-science-center-grand-


46. Leonard de Vinci, homme de science 1452-1519. Paris,
couverte, Dec. 1952 – Jan. 1953. France: Palais de la De 47. U.S. Naturalization Records Indexes, the New York State, Nov. 11, 1954. 48. “A Filadelfia Leonardo inventore,” L’italia, Jan. 18, 1952, Art. no. 3. 49. Mail of J. Mirabella to V. Monnier, 2013 (V. Monnier, personal communication, Jan. 2020). 50. C.-T. Wu, “Privatizing culture,” Verso, pp. 210–211,

opening; C. Kaller, personal communication, Jan. 2020.
nin, “La calcolatrice di Leonardo,” Mondo 56. S. He Digitale, Oct. 2018, Accessed: Feb. 12, 2020. [Online]. Available: http://mondodigitale.aicanet.net/2018-5/ Rubriche/01_MD78_Henin_IlCalcolatoreDiLeonardo.pdf. 57. Eight original Pascaline are known, six of which are distributed in two museums (four at CNAM in Paris, two at Henri-Lecoq Museum, in Clermont-Ferrand,

2002. We also found the following statements:

one is owned by IBM, and another one by an

J. Strickland, “Nathan Myhrvold [former Microsoft

unknown private collector. J. Marguin, Histoire des

manager] bought the remaining calculating replicas”.

 calculer, Hermann, Instruments et machines a

[Online]. Available: http://ed-thelen.org/comp-hist/

1994, p. 62; Arithmetical Machines and Instruments,

hollerith.html; “The [Leonardo’s] replicas are now

Accessed: Jan. 23, 2020. [Online]. Available: http://

owned by the Gallery Association of New York State (GANYS)” (B. Atalay, Math and the Mona Lisa, The Art

ami19.org/. 58. The original is at Gottfried Wilhelm Leibniz Bibliothek –

and Science of Leonardo da Vinci, Smithsonian Book,

€chsische Landesbibliothek, Hanover, Niedersa

2007, pp. 203-204). Still later: “We visited their [IBM’s]

Germany. Accessed: Jan. 23, 2020. [Online].

climate-controlled artifact storage warehouse in Elms-

Available: https://www.gwlb.de/Leibniz/Leibnizarchiv/

ford, New York that housed their pre-computer collection, which included one of the few priceless Pascaline

Leben_und_Werk/. 59. The original is at the Science Museum, London, UK.

calculators outside of France.” (G. Bell, Out of a

Accessed: Jan. 23, 2020. [Online]. Available: https://

Closet: The Early Years of The Computer [x] Museum,

collection.sciencemuseumgroup.org.uk/objects/

Microsoft Corporation, April 4, 2011. Accessed: Feb.

co62243/difference-engine-no-1-difference-engine-

13, 2020. [Online]. Available: https://www. researchgate.net/publication/267569839_Bell_

portion-only. 60. One original is at Endicott History & Heritage Center,

Gordon_Out_of_a_Closet_The_Early_Years_of_The_

40 Washington Ave. Endicott, NY. Accessed: Jan. 23,

Computer_Museum_Dedicated_to_Brian_Randell_on_

2020. [Online]. Available: http://ed-thelen.org/comp-

the_Occasion_of_his_75th_Birthday). 51. IBM Corporate Archives, Accessed: Jan. 23, 2020.

hist/IBM-EndicottMuseum.html. 61. The original is at the Science Museum, London.

[Online]. Available: https://www.ibm.com/ibm/history/

Accessed: Jan. 23, 2020. [Online]. Available: https://

exhibits/index.html.

collection.sciencemuseumgroup.org.uk/objects/

52. M. L. di Vinci, Accessed: Jan. 28, 2020. [Online].

co62245/babbages-analytical-engine-1834-1871-trial-

Available: http://www.museoleonardiano.it/ita/museo/

model-analytical-engines. Another replica was in the

la-storia, Roberta Barsanti, personal communication,

IBM collection (IBM, A Calculator Chronicle. 300 Years

Jan. 2020.

of Counting and Reckoning. IBM, 1990).

April-June 2020

31

Article

62. Georg and Edvard Scheutz made two Difference

72. Arithmeum, Bonn, Germany. Accessed: Feb. 28, 2020.

Engines after Babbage’s. One original is at the

[Online]. Available: https://www.arithmeum.uni-bonn.

National Museum of American History, Washington.

de/en/arithmeum/museum.html; Ina Prinz, personal

Accessed: Jan. 23, 2020. [Online]. Available: https://

communication, Feb. 2020.

americanhistory.si.edu/collections/search/object/

73. C. Randone, Personal communication, Jan 2020.

nmah_997042and the second is at the Science

74. National Museum of American History Collection.

Museum, London, UK.

Accessed: Feb. 28, 2020. [Online]. Available: https://

63. The original was destroyed by Poleni himself. A

americanhistory.si.edu/collections/search/object/

different reconstruction, due to Franco Soresini, was

nmah_904254. In 1957 IBM asked the London Science

built in IBM-Italia workshop in 1959, and is now at the

Museum information about the sizes of the original

Museo Nazionale Scienza e Tecnologia Leonardo da

(R. Cook, Science Museum, London, personal com-

Vinci, Milan, Italy.

munication, Jun. 2012). The Science Museum website

64. H. Bruderer, “The model maker of Leonardo da Vinci,

attributes the Washington replica to Munro Ltd., made

Blaise Pascal, and Charles Babbage,” BLOG@CACM,

in 1957. Accessed: Feb. 28, 2020. [Online]. Available:

Dec. 12, 2018, Accessed: Jan. 12, 2020. [Online].

http://babbagedifferenceengine.blogspot.com/2014/

Available: https://cacm.acm.org/blogs/blog-cacm/

11/fragment-of-babbages-first-difference.html.

233366-the-model-maker-of-leonardo-da-vinci-blaise-

75. B. Korte, Personal communication, Feb. 2020.

pascal-and-charles-babbage/fulltext; H. Bruderer,

76. L. P. Hartley, The Go-Between, County Homes, 1963.

“More replicas of historic calculating machines found,”

77. W. Benjamin, “The work of art in the age of mechanical

BLOG@CACM, Jan. 11, 2019, Accessed: Jan. 12,

reproduction,” in H. Arendt and H. Zorn (eds),

2020. [Online]. Available: https://cacm.acm.org/blogs/

Illuminations, New York, NY, USA: Schocken

blog-cacm/234005-more-replicas-of-historicalcalculating-machines-found/fulltext.

Books, 1969. 78. R. Friedell, “Science and technology in the twentieth

65. Hollerith Census Machine, Accessed: Jan. 23, 2020. [Online]. Available: http://ed-thelen.org/comp-hist/

century exhibitionary complex,” in Canadelli, Beretta, Ronzon6 (2019), cit. pp. 235–247, p. 246.

hollerith.html. 66. B. Catania, “Sulle tracce di Antonio Meucci. Appunti di viaggio”, L’Elettronica, vol. 79, no. 10, 1982, Art. no. 981. 67. J. C. McPherson, F. E. Hamilton, and R. R. Seeber Jr, “A large-scale, general-purpose electronic digital calculator: The SSEC,” IEEE Ann. History Comput., vol. 4, no. 4, pp. 313–326, Oct. 1982. 68. Carnegie

Mellon

University,

Traub-McCorduck

Collection. Accessed: Jan. 23, 2020. [Online]. Available: https://www.library.cmu.edu/about/ publications/news/traub-mccorduck-collection. 69. Computer History Museum, Sunnyvale, CA, USA Accessed: Jan. 23, 2020. [Online]. Available: https:// www.computerhistory.org/collections/search/. 70. Canada Science and Technology Museum Accessed: Jan. 23, 2020. [Online]. Available: https:// ingeniumcanada.org/ingenium/collection-research/ collection.php. 71. Museo Nazionale Scienza e Tecnologia Leonardo da Vinci, Milan, Italy. Accessed: Feb. 28, 2020. [Online].


nin (Member, IEEE) is a scholar of the hisSilvio He tory of computing, author of several books and articles, member of AICA (Associazione Italiana per il Calcolo Automatico e l’Informatica), coordinator of the working group “History of Informatics” of AICA, member of the scientific board of Mondo Digitale, and consultant of Museo Nazionale Scienza e Tecnologia Leonardo da Vinci, Milan, Italy. He is the corresponding author of this article. Contact him at [email protected]. Simona Casonato is curator of Media, ICT, and Digital Culture collections with the Museo Nazionale Scienza e Tecnologia “Leonardo da Vinci,” Milan, Italy. She has been working with the museum since 2003 as audiovisual producer, screenwriter, content manager, collection historian, and curator. She currently collaborates with universities in Italy and abroad for research on the museology of media and computing. Contact her at [email protected].

Available: https://www.museoscienza.org/it/collezioni/ cataloghi-online.

32

IEEE Annals of the History of Computing

Article

The First Computer in New Zealand Brian E. Carpenter The University of Auckland

Abstract—How quickly did the computer revolution reach the most remote Westernized country? Conventional history holds that the first modern computer in New Zealand— where “modern” means electronic, and with stored programs—was an IBM 650 leased from IBM Australia by the New Zealand Treasury in November 1960, and officially inaugurated in March 1961. This article discusses an alternative hypothesis—that the pioneer was in fact an ICT 1201 ordered in 1959 and installed by the New Zealand Department of Education a few months before the arrival of the IBM 650.

& THE MODERN COMPUTER age was announced to New Zealand almost as soon as it began. For example, in September 1949 the Otago Daily Times published an article1 describing the Cambridge EDSAC, the first full-scale electronic computer with an internally stored program ever to operate. At the same time, machinery for financial and other calculations was already widespread in New Zealand government and business. This machinery was largely based on punched cards, and was entirely electro-mechanical, without electronics. As a result it was very noisy, needed skilled operators, and required skilled mechanics to keep its delicate mechanisms in good running order. The New Zealand Treasury, for example, had a busy chief mechanic

Digital Object Identifier 10.1109/MAHC.2020.2990647 Date of publication 27 April 2020; date of current version 29 May 2020.

April-June 2020

very appropriately named Mr. Dainty.2 However, the workload for these machines was constantly increasing. When electronic computers capable of handling punched cards became commercially available in the late 1950s, they were of immediate interest to the Treasury in particular. Treasury installed its first computer, an IBM 650, in 1960 and this is normally recognized as the first modern computer in New Zealand. However, there is an alternative contender.

CONVENTIONAL HISTORY The historical record for the IBM 650 is quite clear. According to A.C. Shailes, who was New Zealand’s Controller and Auditor-General from 1975 to 1983, but an ordinary Treasury official at the relevant time, IBM Australia proposed the lease of an IBM 650 to the New Zealand Treasury in 1957.3 This was a safe choice; well before 1960, the 650 was firmly established in the USA as the

Published by the IEEE Computer Society

1058-6180 ß 2020 IEEE

33

Article

Figure 1. HEC4 Computer (BTM advertising).

bridge from card processing to computing, for example in the insurance industry.4 At that time, it was still IBM’s policy in Australasia only to lease computers, not to sell them. A 650 went for about $3500 to $4000 per month in the USA5,6 but would cost £73000 per year in New Zealand.7 The New Zealand pound was at parity with the British pound at that time, and this was an enormous mark-up by IBM, the exchange rate being fixed at 2.8 $/£. The Treasury moved at normal government speed, so the 650 was finally installed and running at the end of November 1960, and was officially inaugurated on 23 March 1961.7 Its biggest single job was the fortnightly payroll for “some 34 000 public servants,” calculating payments due and deductions, and printing pay vouchers and punched-card format pay cheques for them.3 This machine was phased out in 1968, when IBM donated it to the young Museum of Transport and Technology (MOTAT) in Auckland,7 although the only trace of it in MOTAT’s catalogue is a press photograph taken for the Auckland Star in March 1968.8 Government use of computers grew rapidly after 1960; for example the Department of Education is cited as having its own machine by 1962.7 On the academic side, the New Zealand Computer Society was formed in 1960,3 and universities were early adopters, with the University of Canterbury first in 1962.9 Thus, runs the conventional history of the pioneering years of 1960–1962, but it is incomplete.

DEPARTMENT OF EDUCATION COMPUTER The trigger for this article came from one of the first computer programmers in New Zealand outside government, Ruth Engleback.10,11 Born

34

Ruth Thomson, from 1954 she was trained and employed as a programmer in Stevenage, England by the British Tabulating Machine Co., Ltd., (BTM), leaving in July 1959 for personal reasons. Among other things, she recalls writing a payroll program for 30 000 education personnel (teachers and other school employees) at Middlesex County Council in 1958. By 1959, she was an expert programmer of the BTM 1201, colloquially known as a HEC4 (Hollerith Electronic Computer 4, Figure 1), the commercial derivative of the APE(R)C machine developed by Kathleen and Andrew Booth at Birkbeck College, London.12 In 2012, Ruth recalled hearing before she left BTM that a HEC4 had been sold to a customer in New Zealand. That would place the sale at the latest in the middle of 1959, which does not of course indicate the delivery date. In early 1959, BTM formally amalgamated with Powers-Samas, its main competitor in the British market for punched card equipment, to form International Computers and Tabulators, Ltd., (ICT). This was not without repercussions at the other side of the world. In March 1959, PowersSamas Accounting Machines (Sales), Ltd., of Cambridge Terrace, Wellington, New Zealand announced that it would trade as the PowersSamas group of International Computers and Tabulators, Ltd., from 30 June 1959.13 So the machine in question might be referred to as a HEC4, a BTM 1201, or an ICT 1201, and the local agent, even if officially ICT, might be referred to inaccurately as Powers-Samas, anachronistically as BTM, or colloquially as “Hollerith,” as we shall see below. The delivery records of the BTM/ICT 120114 show two machines delivered to Middlesex County Council, London, in 1958 and 1959, respectively, with the stated purpose “Payroll for 40 000,” closely corresponding to Ruth Engleback’s memory. They also show one machine delivered to the New Zealand Department of Education, Wellington, New Zealand, for the purpose of “Teacher’s salaries.” Unfortunately there is no date in this entry; its position in the list suggests 1960, which is consistent with an order placed in 1959. To take the story further, we must digress from Ruth’s story to look at the general history of punched card equipment in New Zealand Government. New Zealand was far from being a technological backwater, even between the World

IEEE Annals of the History of Computing

Wars, although trailing a few years behind Britain. In 1920, New Zealand had chosen “Powers” equipment for processing the 1921 census data, and rejected the “Hollerith” alternative.15 Apparently this was because the “British Tabulating (Hollerith) Machine Company” declined to sell its machines to New Zealand, insisting on a rental agreement.16 This was required until 1949 by BTM’s licence to use IBM technology.17 A Treasury file “Machines—Powers Samas and Hollerith” covering the period from the 1930s to 19542 shows much more Powers-Samas activity than “Hollerith,” apparently with no mention of BTM as the “Hollerith” manufacturer. The following paragraphs rely on numerous documents in this file. The New Zealand Government commitment to Powers-Samas appears to have started early; a 1953 memo from the Director-General of the Post Office to Treasury refers to some “Powers” equipment in continuous use since 1927. The names “Powers” and “Powers-Samas” were used interchangeably as the company’s formal name evolved.18 An unsigned note dated November 2, 1938 and headed “Office Machinery—Land and Income Tax Department,” describes Tax officials being “surprised to ascertain that there was any other make than Powers” and stating that “prompt and complete servicing was in Treasury view absolutely essential, that this was available here for Powers but was not available and had seemed somewhat doubtful for Hollerith.” Little had changed by June 2, 1949, when a Treasury memo entitled “Punch-Card Accounting and Other Installations–Powers Versus Hollerith,” signed by A. McGregor and sent to the Secretary of the Public Service Commission, said “Hollerith machines have not so far been installed in any Government Department.” They were considered expensive, could only be rented (although in 1949, this was about to change), and the makers “have not been prepared to take a chance and set up an adequate organization in this country.” Even when Powers-Samas switched from 45 to 65 column cards in the 1950s, causing compatibility issues, this did not appear to change Treasury’s preferences. “Hollerith,” following IBM, had used 80 columns from 1928 onwards. Consistently with McGregor’s memo, there is no trace of BTM itself having staff or offices in New Zealand prior to the merger with their

April-June 2020

competitor Powers-Samas, although both their products were imported because of their support for pounds, shillings, and pence. There is ample proof that the term “Hollerith” was familiar in New Zealand by the 1940s, but the name “BTM” was not, although the latest BTM machines were imported, for example by the Reserve Bank.19 The company was well aware of the New Zealand market and appears to have supported it from Australia. For example, in a column headed “Personal Items” in the Wellington Evening Post in June 1939,20 we read that “Mr. A. Stewart Laird, of the British Tabulating Machine Co., London, arrived by the Awatea today on a business visit to New Zealand.” The liner Awatea was on the Australia to New Zealand route from 1936 until she was requisitioned for war service. In fact, Mr. Laird of Sydney appears several times in newspaper lists of noteworthy hotel guests in 1938 and 1939. Thus, before the late 1950s when ICT was formed, it appears that BTM supported its New Zealand customers from its Australian subsidiary, the company still being known colloquially as “Hollerith” on both sides of the Tasman Sea.21 By the end of World War II, government usage of punched card equipment was widespread. In 1945 and 1946, the Public Service Commissioner’s Office advertised repeatedly for “Accounting and Statistical Machine Operators (female),” stating that “experience with “Powers-Samas” or “Hollerith” machines” was unnecessary as training would be given.22 The starting weekly salary for a 16-year-old was £2/1/6, or NZ$4.15. Various items in the Treasury file2 make it clear that punched card operations were considered an exclusively female occupation and that the managers were men. These women have left little trace in the public record of their role in computerization of their work. There is no reason to suppose that they were less able than the women of ENIAC23 or any others involved in data processing at the time.24,25 Returning to the needs of the Department of Education, it already had a Burroughs ledger machine for superannuation accounts before the end of World War II, according to a letter from G. V. Brooke in the Burroughs office in Wellington to the Director of Education, dated June 27, 1945.2 From about 1950, the Department also borrowed time on punched card equipment

35

Article

from other Government Departments for statistical work. However, in 1954, the Department decided to obtain its own punched card equipment, because of expected changes in teachers’ salary payments. On July 16, 1954, the Director of Education wrote to the Treasury about a longstanding request from the teachers’ union for fortnightly, instead of monthly, pay. This had been deemed impossible “in view of the large cost that will be involved if the system is introduced under the present methods operating in each Education Board.” The memo recommended a central punched card solution. Furthermore, the memo stated. Approaches were made, without any commitment, to the two leading firms in Wellington dealing with punched card installation and each has supplied a brochure setting out a system that could do the work we want and an estimate of the cost of installation. . . . The full installation is expected to cost approximately £20 000. . .

One of these firms was undoubtedly the Powers-Samas office in Wellington, and the other must have been whoever represented BTM locally in 1954. A memo from the Secretary to the Treasury, D. Barker, to the Minister of Finance dated August 4, 1954 referred to the proposal to move teachers from monthly to fortnightly pay, requiring the “installation of additional punched card machinery costing up to £18 500.” It also noted that the scheme “could not be implemented before 1956,” an optimistic estimate as things turned out. The memo also observed that Treasury used punched card equipment to handle the payroll for 20 000 public servants in 26 departments, who were already paid fortnightly. On August 13, 1954, Barker sent a memo to the Director of Education approving a modest initial purchase of PowersSamas punched card equipment by the Department of Education, to be ordered via the NZ High Commission in London. In the same memo, Barker referred to proposals from both PowersSamas and “Hollerith” for “the major scheme” and agreed that “a central machine plant is likely to provide the most economical and efficient system.” Previously, teachers’ pay was administered by various Education Boards around the country; the new scheme envisaged “preparing all cheques at the central point.” Centralization of the

36

payroll was part of a long process of weakening of the regional Education Boards, which were finally abolished by the Education Act of 1989.26 However, there is no evidence that mechanization of the payroll had any impact beyond the clerical staff directly affected. After almost five years delay, in April 1959, teachers in New Zealand switched from monthly to fortnightly pay cheques. As their trade union’s journal editor wrote in May 1959.27 Having now had the experience of two fortnightly salary payments, teachers are in a position to realise the convenience of the new arrangement as compared with monthly payments. . . . To hold to the Department’s intention of bringing in fortnightly pay in April meant far more work than was at first expected. It meant working back for long hours at night, sometimes well past midnight, and added to this was the mental strain of evolving a new system that included the use of unfamiliar and intricate machinery.

We can only speculate about what machinery this refers to. However, changing from monthly to fortnightly payments was not trivial, and the final decision to do so was taken only in August 195828 for implementation on April 1, 1959.29 First, the payroll needed to be run 26 times a year instead of 12. Second, in the New Zealand currency of the time—pounds, shillings, and pence—any annual salary or allowance that was a whole number of pounds and shillings could be divided by 12 without any rounding errors, so the exact calculation of monthly pay was straightforward. By contrast, the fortnightly calculation would rarely be exact and rounding errors would need to be reconciled. Third, the payroll dates would slip back by one day a year, and by two days in 1960 (a leap year), again leading to tricky coding. It is easy to see why the Department of Education needed new equipment and skills in 1959. In September 1960, the union journal wrote (for once, not complaining about salaries).30 It is anomalous that when teachers’ pay and deductions are processed by electronic bookkeeping, the methods of recruiting teachers have not changed much since the horse-and-buggy days.

This implies that the Department of Education was then using an electronic machine of

IEEE Annals of the History of Computing

some kind to run the payroll every two weeks. Since we also know that they probably acquired their ICT 1201 at some time in 1960, the conclusion is tempting that this was the provider of “electronic bookkeeping” referred to, at least two months before the commissioning of the Treasury’s IBM 650. But this is not quite a certainty. BTM did have electronic products prior to the HEC4 computer: electronic multipliers were included in some of its calculators, including the BTM 555 introduced in 1952.31 Powers-Samas failed to enter the stored-program computer market. In the mid-1950s, they did however produce an electronic device called the PCC (Programme Controlled Computer), which included a magnetic drum memory, but it was not a stored-program computer in the Turing or von Neumann sense. “The programme instructions are not entered into the computing store, but are set up on switches preparatory to a computation taking place.”32 Much of its electronics was concerned with calculations in pounds, shillings, and pence. Although manually programmed, the PCC was quite successful. By 1958, some 25 had been delivered, with overseas orders including Australia.33 Indeed, any machine that could multiply and divide pounds, shillings, and pence could find a market in Australia and New Zealand in those days, as well as in the UK. New Zealand converted to decimal currency only in 1967, one year after Australia and four years before Britain and Ireland. South Africa had converted in 1961. (All were a century behind Canada.) Since PowersSamas had a sales office in Wellington, and the PCC was sold from 1956 onwards, it is certainly not impossible that the Department of Education had a PCC as part of its card-processing facility by 1959. Equally, they could have had a BTM 555 as part of a “Hollerith” installation. However, there is no evidence of either, whereas the ICT 1201 is a certainty. Fortunately, we now know what the vendor thought about the race with IBM. Bruce McMillan joined Powers-Samas as a technician in Dunedin in 1958, automatically becoming an ICT employee in 1959. He attended courses at the Wellington office, and vividly recalls the annoyance of “the Powers/ICT/ICL people about IBM’s claim to be first.”34

April-June 2020

The balance of probability is thus that the Education machine was installed and running by August 1960, but that the Treasury and IBM either did not know this or preferred to ignore it when inaugurating the IBM 650 in March 1961. Another uncertainty is how the Department of Education paid for its ICT 1201. The Department issued an annual report to Parliament in the name of the Minister of Education.35 Surprisingly, the reports for the relevant years appear to ignore the change to fortnightly pay and the associated centralization of payroll handling. The reports do include a reasonably detailed financial report until the fiscal year 1960–1961, but there is no line item that could readily be interpreted as a computer purchase. However, starting in 1957 there are two relevant items: “Hire of punch card machines” and “Duty and tax on accounting machines”; for details see the following table, which also shows a coincidental tendency to increase in “Office equipment.” Budget year

Hire of punch card machines

Duty and tax on accounting machines

Office equipment

1955–1956

£0

£0

£11432

1956–1957

£0

£0

£14178

1957–1958

£1560

£1170

£19600

1958–1959

£4959

£1409

£15300

1959–1960

£10510

£2813

£15764

1960–1961

£15954

£7922

£24300

1961–1962

n/a

n/a

n/a

A reasonable conclusion, considering the likely price of up to £40 0006 for an ICT 1201, is that the Department rented the machine from “Hollerith” (by then formally known as ICT, and established in the old Powers-Samas office in Wellington). Alternatively, perhaps they arranged some sort of lease-to-buy contract. It is tempting to assume that they embedded it in the designation “punch card machines” to avoid any difficulty with the Treasury officials who had approved a punched card project in 1954. In any case, the increases in expenditure seem to match the installation of a large machine by 1960. The Department of Education acquired its second computer, an ICT 1301, in 1962, although

37

Article

conventional history7 gives this as their first machine. The ICT deliveries list for the 130136 shows a machine delivered there for “Teachers’ salaries, school certificate examination results, stores control.” Despite its similar identifier, the 1301 was completely incompatible with the 1201, being a new design manufactured by GEC.6 It is unlikely that the Department would have wanted to keep both, after the new machine was running well, inevitably with completely rewritten software. By late 1963, Ruth Engleback had emigrated to Auckland, New Zealand, and was bringing up a young family. Through a contact of her husband’s, she was head-hunted by a local company called Motor Specialties (sic), which supplied components to the motor trade. The US spelling was intentional. (In 1987, Motor Specialties, Ltd., changed its name to Repco Merchants, Ltd.37 The Australian company Repco acquired its first New Zealand store in 1981, apparently by taking over Motor Specialties.)38 In 1963, Ruth’s job for Motor Specialties was setting up their newly acquired ICT 1201 and getting it operating. This took about a year. Although not the nub of this article, it is worth observing that in 1963 Motor Specialties was, by New Zealand standards, a very early adopter of electronic computing for nongovernmental purposes. Probably the first commercial user was Griffin and Sons, the confectioner, in 1961.7 The Bank of New Zealand (a commercial bank, but state-owned at that time) first called for tenders for two computers in 1964, in anticipation of decimalization three years later.39 The Cadbury chocolate factory in Dunedin bought an ICT 1301, as did Electronic Data Processing, Ltd., in Auckland.36 The latter company is something of a phantom. According to the New Zealand Gazette,40 it was formed in 1961 and struck off the companies register in late 1963, but it had no entry in the 1962 Auckland telephone directory. Cadbury received their machine in November 1963; an earlier ICT 1301 was delivered to Cadbury in Tasmania during 1962.21 The Dunedin machine is now to be seen in the Toitu Otago Settlers Museum in Dunedin after being used, almost incredibly, until 1975.41 Components of a BTM 555 electronic calculator and various PowersSamas machines are also displayed there.

38

Figure 2. Motor Specialties building (Auckland Museum).

In any case, in 1963 Motor Specialties became one of the first commercial users to install a computer in New Zealand, and the first one known to install a second-hand machine. John Robb, who worked for Motor Specialties in the 1960s, recalls that first computer being installed in their premises at 80 Anzac Avenue in Auckland (Figure 2) in 1963.42 The handsome building, designed in 1929 by the Dunedin Artsand-Crafts architect Basil Hooper, still stands in 2020, unsympathetically modernized and home to two language schools. The intended application for the ICT 1201 was management of the company’s large inventory of spare parts for motor vehicles. Robb recalls that the results were not very satisfactory and that the computer was replaced within a few years. The company eventually became highly dependent on a computerized inventory, as its business grew constantly throughout the 1960s—the number of licensed motor vehicles in the country grew by 39% during the decade.43 Neither Motor Specialties, nor any other customer in Auckland, features in ICT’s delivery list for the 1201. Ruth’s recollection is that the computer was bought second-hand from a government department and moved up from Wellington, at a cost of about £15 000. Since the price for a new machine was up to £40 000,6 this is a plausible second-hand price for a machine that was presumably considered obsolete by the Department of Education, the only possible source in New Zealand of a used ICT 1201 in 1963. If, as conjectured above, the Department had actually rented the machine, Motor Specialties probably bought

IEEE Annals of the History of Computing

it directly from ICT. Even so, the available data make the machine seem like an excellent bargain for the taxpayers, compared to the exorbitant rental fee for the Treasury’s IBM 650. Ruth further recalled that the 1201 was eventually donated by Motor Specialties to the Auckland Technical Institute (ATI), which later became Auckland University of Technology (AUT). The formal donation is well documented and took place on December 15, 1966.44,45 She was called in again to train ATI staff to program the 1201. She recalls that Motor Specialties had by then acquired an ICT 1301, which is not listed in the ICT records.36 It is possible that this was the machine sold to the short-lived Electronic Data Processing, Ltd. Motor Specialties remained loyal to ICT and ICL for almost 20 years, before finally becoming an IBM shop. Ruth’s final recollection is that ATI donated the ICT 1201 to MOTAT. Unfortunately, neither AUT nor MOTAT has found a record of this. It should be noted that MOTAT was founded only in 1964, by aeronautical and transport enthusiasts, who probably did not discern the importance of early computers such as the ICT 1201 and the IBM 650, and did not take cataloguing as seriously as today’s professional staff, who diligently searched their archives in 2019.

CONCLUSION We have shown that the New Zealand Department of Education ordered an ICT 1201 computer in 1959, that it was in all probability operational by August 1960, and that its supplier believed it to be first in the NZ market. There is thus a strong likelihood that the first operational stored-program electronic computer in the country was not an IBM 650, but an ICT 1201. This machine later became the first second-hand computer installed by a commercial user in New Zealand. Why was the early acquisition of an ICT 1201 so little known? Even an oral history project covering 1960 to 2010 failed to dislodge the Treasury’s claim to precedence.46 We can only speculate, but it could be that, aware of the Treasury’s slow-moving acquisition of the IBM device, the Department of Education considered it best to fly under the radar and simply acquire equipment that it desperately needed for its

April-June 2020

fortnightly payroll without publicity. 25 years later, when the NZ Computer Society celebrated its jubilee, the machine was not even mentioned. In the Treasury files,2 it is very noticeable that Treasury officials considered themselves to be in charge of all expenditure. Foreign exchange was tightly controlled by the Reserve Bank, import licences were needed for almost everything, import duty was high and the economy was highly regulated.47 In short, computers were expensive. All these are possible reasons why the Department of Education avoided publicity that might alarm the Treasury. This not only had a perverse effect on historians by effectively concealing the computer. Although universities were early adopters, computing was relatively slow to enter school curricula in New Zealand.48 If the Department of Education had been a recognized leader, the schools might have taken notice much sooner. Finally, we should note the role of women in this story. Ruth Engleback was far from alone in the pioneering days, and she, her peers, and the many women who operated punched card equipment before and after computerization deserve their place in history, from which they have largely been excluded until recently.24,25 The end of 1960 was a decisive moment: mechanical data processing was coming to an end, new suppliers and products were appearing, the programming of machines was changing from physical labor to desk work, and thousands of clerical jobs were about to be transformed. The New Zealand Department of Education should share the credit for leading the way.

ACKNOWLEDGMENTS Many thanks are due to Ruth Engleback for her role in triggering this investigation, to my late colleague Robert W. Doran for starting the work, and to the anonymous reviewers. Staff at the following institutions were very helpful: the Museum of Transport and Technology (MOTAT) in Auckland, Archives New Zealand, the Alexander Turnbull Library at the National Library of New Zealand, the Auckland University of Technology Library (Special Collections), the University of Auckland Library, the Auckland Central City Library, and Toitu Otago Settlers Museum. Valuable help in tracing UK references came

39

Article

from Martin Campbell-Kelly and Roger Johnson. Special thanks go to the NZ Government on-line archives at https://paperspast.natlib.govt.nz and to those of NZEI Te Riu Roa at https://heritage. nzei.org.nz.

15. Machines and the Census (Report of a Lecture by

& REFERENCES

17. M. Campbell-Kelly, “ICL company research and

Mr. Malcolm Fraser, the Government Statistician), Auckland Star, Jun. 7, 1921. 16. M. Fraser, “Report of government statistician on British empire statistical conference,” Appendix J. House Representatives, 1920 Session I, H-12. development, 1904-1959,” ICL Techn. J., vol. 5, no. 1,

1. Electronic Brain—Remarkable Machine—Installation at Cambridge University, Otago Daily Times, Sep. 3, 1949. 2. Machines—Powers Samas and Hollerith, Archives New Zealand Record R15421250, 1954.

Feb. 3, 2020. [Online]. Available: https://www. gracesguide.co.uk/Powers-Samas_Accounting_

3. A.C. Shailes, The Impact of Computers on the Public Sector, Chapter in Looking Back to Tomorrow,

Machines 19. Reserve Bank Museum, Wellington, Photograph of

W. R. Williams, Ed. Auckland, New Zealand: New

Installation of Hollerith Machine (Rolling Total

Zealand Comput. Soc., 1985, pp. 35–52.

Tabulator), 1948, Accessed on: Dec. 28, 2019.

4. J. Yates, Structuring the Information Age: Life Insurance and Technology in the Twentieth Century. Baltimore, MD, USA: Johns Hopkins Univ. Press, 2005. 5. P. E. Ceruzzi, A History of Modern Computing. Cambridge, MA, USA: MIT Press, 1998.

[Online]. Available: https://www.nzmuseums.co.nz/ collections/3369/objects/2680/installation-of-hollerithmachine-1948 20. Personal Items, Evening Post. Wellington, New Zealand, Jun. 10, 1939.

6. J. Hendry, Innovating for Failure: Government Policy

21. R.G. Bird, Oral History Interview by Thomas J. Misa on

and the Early British Computer Industry. Cambridge,

19 November 2013. Charles Babbage Institute, 2013,

MA, USA: MIT Press, 1989.

Accessed on: Dec. 28, 2019. [Online]. Available:

7. C. Beardon, Computer Culture—The Information Revolution in New Zealand. Auckland, New Zealand: Reed Methuen, 1985, p. 7. Star, Accession No. 04-4245, Walsh Memorial Library, The Museum of Transport and Technology (MOTAT), Mar. 1968, Accessed on: Feb. 25, 2020. [Online]. Available: https://collection.motat.org.nz/objects/30577 9. T. Dale, Early Computing at the University of Canterbury. Dec. 2015, Accessed on: Apr. 16, 2020. [Online]. Available: https://wiki.canterbury.ac.nz/ download/attachments/49417267/cabinetcontents.pdf 10. R.

W.

Doran,

New

Zealand’s

https://conservancy.umn.edu/handle/11299/164967 22. Situations Vacant, Evening Post. Wellington, New Zealand, Nov. 28, 1945.

8. IBM 650 computer, press photograph, The Auckland

First

Computer

Programmer, Blog Entry About Ruth Engleback, 2012, Accessed on: Dec. 27, 2019. [Online]. Available: https://universal-machine.blogspot.com/2012/11/new-

23. W. B. Fritz, “The women of ENIAC,” IEEE Ann. History Comput., vol. 18, no. 3, pp. 13–28, 1996. 24. T. Haigh, “Masculinity and the machine man: Gender in the history of data processing,” in Gender Codes: Why Women Are Leaving Computing, T. J. Misa Ed. Hoboken, NJ, USA: Wiley, 2010. 25. J. Abbate, Recoding Gender: Women’s Changing Participation in Computing. Cambridge, MA, USA: MIT Press, 2012. 26. The government’s changing role in the governance of New Zealand’s schools since 1847, NZ Parliamentary Service, Dec. 2019. 27. National Education—The Journal of the New Zealand

zealands-first-computer-programmer.html

Educational Institute, vol. XLI, no. 443, May 1959.

11. R. Engleback, Personal communication, 2019.

28. National Education—The Journal of the New Zealand

12. M. Campbell-Kelly, ICL—A Business and Technical History. Oxford, U.K.: Clarendon, 1989. 13. The New Zealand Gazette No. 17, Mar. 19, 1959. 14. Delivery Lists and Applications of the BTM HEC Computers and the BTM / ICT 1200 Series, Computer Conservation Society, 2011, Accessed on: Dec. 27,

40

pp. 2–17, 1986. 18. Powers-Samas Accounting Machines, Accessed on:

Educational Institute, vol. XL, no. 436, Sep. 1958. 29. National Education—The Journal of the New Zealand Educational Institute, vol. XL, no. 437, Oct. 1958. 30. National Education—The Journal of the New Zealand Educational Institute, vol. XLII, no. 458, Sep. 1960. 31. Raymond Bird, “BTM’s first steps into computing,”

2019. [Online]. Available: http://www.

Comput. Resurrection Bull. Comput. Conservation

ourcomputerheritage.org/ict_15.htm

Soc., vol. 22, pp. 12–18, 1999.

IEEE Annals of the History of Computing

32. E. J. Guttridge and R. P. B. Yandell, “The programmecontrolled computer—A digital computer for

41. Dunedin’s first computer cutting-edge in ’63, Otago Daily Times, 7 Sep. 2013.

commercial applications,” Proc. IEE—Part B: Radio

42. J. Robb, Personal communication, Jan. 10, 2020.

Electron. Eng., vol. 103, no. 2, pp. 217–227, 1956.

43. New Zealand Official Year Books, Statistics New

33. N. Calder, “The development and uses of digital computers,” New Scientist, vol. 4, no. 106, pp. 1383–1386, Nov. 27, 1958. 34. B. McMillan, personal communication, Feb. 2020. 35. Reports of the Minister of Education for the Year Ended 31st December, 1955 through 1961, Appendices to the Journals of the House of Representatives, Wellington, New Zealand. 36. BTM/ICT 1301 Delivery List, Computer Conservation Society, 2004, Accessed on: Dec. 27, 2019. [Online]. Available: http://www.ourcomputerheritage.org/ict_15. htm

Zealand, 1961 and 1971. 44. J. F. Johnston, Letter from J.F. Johnston, Managing Director, Motor Specialties, to R. A. Keir, Principal, Auckland Technical Institute, Nov. 18, 1966. 45. R. A. Keir, Letter from R.A. Keir to J.F. Johnston, Nov. 23, 1966. 46. J. Toland and J. Whitman, Wellington’s Computer Pioneers 1960 to 2010, Oral History in New Zealand, vol. 27, pp. 1–8, 2015. 47. The Reserve Bank and New Zealand’s Economic History, Reserve Bank of New Zealand, 2007. 48. D. Ferguson, “Development of technology education in

37. The New Zealand Gazette No. 90, Jun. 18, 1987.

New Zealand schools 1985–2008,” NZ Ministry Educ.,

38. Repco, Ltd. Accessed on: Dec. 28, 2019. [Online].

Sep. 2009.

Available: https://www.repco.co.nz/en/about-us 39. I. H. Archibald, Computers and Banking, Chapter in Looking Back to Tomorrow. W. R. Williams, Ed. Auckland, New Zealand: New Zealand Comput. Soc., 1985, pp. 53–75. 40. The New Zealand Gazette No. 65, Oct. 24, 1963.

April-June 2020

Brian E. Carpenter is currently an Honorary Professor of computer science with The University of Auckland, Auckland, New Zealand. His research interests include Internet protocols, especially the networking and routing layers, as well as computing history. Contact him at [email protected].

41

Article

Once FITS, Always FITS? Astronomical Infrastructure in Transition Michael Scroggins University of California

Bernadette M. Boscoe University of Washington

Abstract—The flexible interchange transport system (FITS) file format has become the de facto standard for sharing, analyzing, and archiving astronomical data over the last four decades. FITS was adopted by astronomers in the early 1980s to overcome incompatibilities between operating systems. On the back of FITS’ success, astronomical data became both backward compatible and easily shareable. However, new advances in the astronomical instrumentation, computational technologies, and analytic techniques have resulted in new data that do not work well within the traditional FITS format. Tensions have arisen between the desire to update the format to meet new analytic challenges and adherence to the original edict for the FITS file format to be backward compatible. We examine three inflection points in the governance of FITS: first, initial development and success, second, widespread acceptance and governance by the working group, and third, the challenges to FITS in a new era of increasing data and computational complexity within astronomy.

CAN A FILE FORMAT GOVERN? I may be blasphemous here [on the Vatican adopting FITS], I [think], "Oh that’s another sign that things are in bad shape”.. . . Maybe we’re being a little too conservative here – Astronomer1

Digital Object Identifier 10.1109/MAHC.2020.2986745 Date of publication 13 April 2020; date of current version 29 May 2020.

42

1058-6180 ß 2020 IEEE

In the late 1970s, a handful of astronomers built a file format specifically designed for astronomical data. There is nothing unusual about this in and of itself; scientists commonly built formats, readers, and other low-level tools to do their science. However, the file format they designed, the flexible interchange transport system (FITS), was given an explcit policy edict, FITS was to remain forever backward

Published by the IEEE Computer Society

IEEE Annals of the History of Computing

compatible. The edict, encapsulated in Don Well’s pithy phrase “once FITS, always FITS” is a modern formulation of a tradition borne of the demands and hard experience of observational science; rare events observed at great difficulty are to be kept accessible for future astronomers. This rule enabled FITS to be used for a dizzingy array of astronomical data. Inspired by FITS’ long-term stability, in 2014 the Vatican decided to adopt the format for their image archives.2 Yet, a stable archival format is not necessarily what all astronomers would like; some want a format that has adapted to modern computing capabilities, of which FITS lacks. FITS was formally described in 19813 and by 1982 became the de facto standard for sharing, analyzing, and archiving astronomical data. Widely adopted by astronomical observatories and scholarly associations shortly thereafter because of its ability to overcome the noninteroperability of contemporary computer systems. FITS was an immediate success and within a few short years, working astronomers were able to exchange data with colleagues regardless of which computer systems they used or whether their images were generated via radio or optical telescope. By 1990, when NASA’s Hubble telescope reached first light and the nascent observations were disseminated via FITS files, FITS had become the lingua franca of astronomy, as unavoidable within the confines of astronomy as English is within the broad expanse of science.4 FITS has changed rarely, and then only in small increments. Since 1993, only seven modifications, in total, have been made to the format5 and only five official modifications to the format standard have been made. In keeping with astronomical tradition, which gives the observer control of the mechanics and tooling of the telescope used for observation, FITS remains “Flexible,” open to local modifications and workarounds. Accompanying FITS, and coterminous with its wide adoption, came governance through a loose confederation of scholarly associations with overlapping and interlocking memberships. Between 1988 and 2016, FITS was governed by an official International Astronomy Union (IAU) working committee, the FITS Working Group (FWG), before being superseded by the “Data

April-June 2020

Representation Working Group (DRWG) (p. 2)”.6 As an international organization, the IAU FWG encompasses numerous national and supranational astronomical organizations, such as the American Astronomical Working Group on Astronomical Software (WGAS) and the European FITS Committee. Meetings are held on an ad hoc basis as comments, objections, queries, and suggested changes to FITS from the polity of working astronomers filter up to the FWG from conferences, such as the Astronomical Data Analysis Software and Systems (ADASS) Birds of a Feather (BoF) sessions, feedback from newsgroups, e-mail exploders, and informal conversations between colleagues.7 Governance entered the social science lexicon in the late 1980s as a way to account for the decoupling of policy from the traditional organs of government.8 While academic governance has long been a topic of scholarly concern, scholars have more recently turned their attention to the governance of open source software.9 FITS is unusual in being a publically shared, noncommercial format governed by scholarly associations. We take up FITS’ governance as a form of material politics through the lens of infrastructure. We examine how the governance of FITS is inscribed in technical standards, astronomical institutions, and in the everyday use of FITS, and consider how FITS is made durable by this infrastructure and how, in turn, infrastructure has made FITS’ place in astronomy durable. To this end, we highlight two seemingly paradoxical qualites of FITS within the broader framework of astronomical governance. In section five we discuss how FITS, despite its age and inflexibility, is not an endpoint, but rather a point of departure for ongoing changes and challenges to existing governance. In section six, we illustrate a brewing problem for astronomical governance, demonstrating how FITS technical obsolence has, ironically, only deepened its use and utility. Infrastructure, as Larkin (p. 328)10 argues, constitutes an “architecture for circulation.” In contrast to accounts that assume infrastructure is inevitably invisible until breakdown and failure brings it to our attention,11 Larkin observes that infrastructral objects have valences (along multiple axes) that resist generalizing claims. In

43

Article

astronomy, for instance, telescopes and observatories enjoy both pride of place and pride of funding, whereas objects, such as FITS labor along in relative anonymity. Yet, both are necessary to, and neither alone sufficient for, an astronomical infrastructure. Infrastructures have a life, a biography; a beginning and an end.12 Infrastructure is, therefore, an example of what STS (Science and Technology Studies) theorists have called material politics.13 Through material politics, objects both extraordinary, such as telescopes, and mundane, such as FITS, make social relations durable and persistent through time and across space (p. 186).14 We intentionally focus on the most mundane of astronomical infrastructure. Following Sterne and Russel,15,16 we argue that as a de facto standard, FITS makes meaning within astronomy and examine the ideological stakes of debates over FITS governance and future in astronomy. In the four sections that follow, we first recount the conditions of FITS’ invention and diffusion within astronomy, second, examine the material politics and infrastructure of FITS governance, third, discuss the reappearance of the interchange problem within astronomy, and fourth, conclude with observations on the governance and material politics of emerging forms of astronomical infrastructure.

DON WELLS’ VISION OF AN ASTRONOMICAL INTERCHANGE FORMAT A unique interchange format needs to be very flexible. . . It should provide a mechanism for transmitting any auxiliary parameters that are associated with the image, even though not all these parameters, nor even the nature of all these parameters, can be specified a priori. – Wells et al. 1981.

From the beginning of astronomy, detailed metadata chronicling positions, time, and conditions have accompanied observations. The earliest form of astronomical observations were drawings. Beginning in the mid-19th century, optical plates slowly replaced drawing as the observation format of record. In the middle of the 20th century advances in detecting radio waves made another form of astronomical

44

imaging possible. Growing out of research dating from the 1930s, radio astronomy measures radio waves emitted by celestial objects. It languished until computers powerful enough to handle the Fourier transform inversions necessary to produce images from radio waves came into common use in the late 1960s. When the first radio telescope sky surveys appeared, comparing them to existing optical surveys required a lengthy and error-prone process of fitting images together through trial and error.17 By the late 1970s astronomy was faced with two observational paradigms–optical and radio–developing along divergent paths. Meanwhile, the computational environment of the 1970s was a collage of rival manufacturers and proprietary operating systems. On the cusp of the PC era, minicomputers and mainframes dominated the academic computing landscape: mainframes from IBM, Burroughs & UNIVAC, NCR, Control Data Corporation, Honeywell, Cray Research, Digital Equipment Corporation, Hewlett–Packard, Amdahl Corporation, and International Computers Limited; minicomputers from Digital Equipment Corporation, Wang Laboratories, Data General, Apollo Computer, and Prime Computer.18 This babel of manufacturers and incompatible physical file formats meant that sharing data was onerous, at best. Yet before the philosophical problem of incompatible machines could be faced, the logistical problem of transporting observational data from telescope to home institution had to be overcome. In the late 1970s, it was common for observatory computing systems to be incompatible with university computing systems. After the long trek off the mountain with a physical box containing tapes and floppy disks, astronomers were then faced with the problem of making their data available to themselves for analysis. Out of this chaotic environment, FITS emerged through a serious of informal conversations between astronomers seeking to share data. For FITS, the most important of these conversations occurred in 1976, when Ron Harten, a radio astronomer from the Netherlands visited Donald Wells, an optical astronomer, at Kitts Peak Observatory, and began an ongoing dialogue about the possibility of writing an interchange standard for

IEEE Annals of the History of Computing

astronomy (p. 251).19 From their conversation, Wells later derived the three mandates that characterize FITS: enable radio and optical images to be combined into a single file, be governed through international committees, and remain forever backward compatible.20 Several key design decisions were also made at this time, the most controversial involving the file header. The header file was to contain human readable info, such as when and with what instrument an image was taken. Wells wanted a flexible file header, rather than fixed fields, to accomplish two things: first, allow FITS to be used by the widest possible audience of astronomers, and second, give knowledgeable programmers enough latitude to implement customized solutions. Because they wanted to reach compromise within the polity of working astronomers, Wells and Harten, joined by the radio astronomer Eric Greisen, prototyped FITS using an IBM 360 for the radio data and a CDC 6400 for the optical data, two systems known to be incompatible. When interchange between radio images generated on an IBM 360 and optical images generated on a CDC 6400 was successful, the foundation for the digitization of astronomy had been laid. Although the technical aspects of FITS were largely worked out between Wells and Harten, the problem of generating buy-in and support from the larger polity of working astronomers remained. In short order, through Greisen, FITS gained support from National Radio Astronomy Observatory (NRAO), and backing from NASA. With NRAO and NASA suppportive of FITS, the next logical step was to move governance of FITS from an informal, ad hoc, conversational basis to a structure with formalized decision making and institutional support. In this effort, FITS was helped by astronomy’s long history of internationalism. The logical institution to lead the governance of FITS was the International Astronomy Union (IAU). The IAU was formed in 1919 to further international cooperation and communication in the wake of the Great War.21 Using the IAU’s supranational position within astronomy and wellestablished system of hierarchical working groups gave FITS a traditional, durable, and far reaching infrastructure for governance. At the 1982 IAU

April-June 2020

General Assembly meeting in Patras, Greece, Commission 5 (Documentation and Astronomical Data) adopted the following resolution (Figure 1).

Figure 1. Proposal to the XVlllth General Assembly of the International Astronomical Union. (p. 16).22

The two papers referenced in the resolution constitute first, the initial definition of the FITS in 1981 and second, the initial extension to the FITS format for handling random groups.23 The years between 1982 and 1990 saw the steady adoption of FITS for everyday work in astronomy, led by a steady stream of software for working with FITS files and the adoption of FITS as an output format by major observatories.24 In 1990, NASA formally adopted FITS as the standard format for all NASA-funded astrophysics projects. Also in 1990, the American Astronomical Society (henceforth AAS) WGAS FITS committee was established with Don Wells at its head.25 In support of this new committee, Wells created a listserv to solicit ideas and encourage discussion within astronomy about FITS. Another important venue for soliciting ideas and discussion about FITS established at this time was the BoF group at the newly formed ADASS conference. Because of its tight focus on astronomical computation and cross-over with IAU and WGAS members, it quickly became the venue of choice for prototyping modifications and extensions to FITS. With the institutional support afforded by NASA’s adoption of FITS, the creation of the listserv in 1990, and the initial ADASS conference in 1991, the infrastructure of FITS governance was in place. FITS would be governed by the hierarchical style of academic governance–conferences, committees, working groups, and international associations–filtering suggestions for modifications and improvement upward to the supranational IAU FWG. Before continuing, we must clarify a few specific technical features of FITS. The workflow

45

Article

imagined by Wells in his 1981 framing paper on FITS was a linguistic process of translation into FITS from institution A’s internal format then reverse translation from FITS into institution B’s internal format. Being a grammar for interchange and interoperability, FITS attempts to sidestep the thorny problem of semantics, the meaning and interpretation of data contained within the file.26 But one area where FITS blurs the line between acting strictly as a grammar for interchange and semantics is in the file header.27 The FITS header is taken directly from the ANSI FORTRAN 1977 standard for list input and takes the form of: keyword ¼ value/comment. The FITS header is made of card images, each card taking 80 columns (bytes), originating from systems using 80 column punchcards. A keyword is any 8character ASCII string. The value field conforms to the FORTRAN standard. The comment is a human readable textual annotation on the intention, meaning, and use of the keyword/value pair. The header is where Wells et al. placed the responsibility for explaining parameters that cannot be specified a priori and where the syntactical and semantics elements of FITS become blurred. As Wells et al. observed of FITS headers “coordinate information and auxiliary parameters are important for the unambiguous interpretation of the digital image, particularly when the object of exchange is the intercomparison of sources as seen by various detector systems” (p. 365).3 The lacunae formed in the FITS header system would turn FITS into a format of dialects and creoles.28 All can speak to each other but require a great deal of translation. On a syntactic level, FITS has successfully bridged the differences between optical and radio astronomy and tamed the chaos in computing hardware and software. But delegating the responsibility of explaining the semantic elements of a file, a task requiring contributory expertise in astronomy,29 to the header had the effect of placing a limit on the kind of intra-astronomical interchanges FITS could accomplish. In a later section, we will take up how the interpretation of digital images within astronomy became problematic when the World Coordinate System (henceforth WCS) was introduced into FITS and more recently in the development of complex analytical pipelines within astronomy.

46

FITS AND THE CIRCULATION OF ASTRONOMICAL OBSERVATIONS My intention in forming the newsgroup is that _anyone_, _anywhere_ is welcome to discuss FITS in this place, and that that person can expect that knowledgeable FITS people will be listening. So, if you have something you want to say, go ahead. – Don Wells30

By 1991, the pressing issues FITS addressed in the 1970s–file interchange, in particular–acquired new answers as computing was changed by the network structure of the internet. The computing ground had started to shift as the problem FITS was designed to solve faded into the past, and new problems, and potential new answers, came to the fore. The internet itself would soon evolve from a space of occasionally comical and often frustrating misunderstandings into a fully formed digital infrastructure capable of coordinating far flung collaborators and supplementing the scholarly societies and institutions that had sped FITS diffusion through astronomy in the 1980s.

Here Comes _Everyone_,_Everywhere_ As the internet gained wider adoption in the late 1980s and early 1990s, the impetus to share progress and thoughts about FITS on the internet grew. There were many options. Usenet news groups were a public way to post and discuss news that anyone could join. Listservs were a less public method of sharing information; via an email server, pre-approved listserv members could create posts and exchange information via threads. Augmenting listservs were email exploders, programs that allowed an email sent to a particular address to be forwarded or exploded to multiple email addresses. Wells started with a news group and eventually linked it via email exploder to a listserv, pushing FITS communications to both the public via newsgroup and to listserv members via email. For sake of brevity, and the fact that the content of the newsgroup, associated listservs and the fitsbits email exploder were closely interrelated, we will take liberties and call the intertwined messaging systems a listserv. Listservs became popular with scientific groups in the early 1990s, as both an informal and formal medium, a way to banter and to coordinate formal actions. For FITS, the listserv bridged the space between hallway conversations in

IEEE Annals of the History of Computing

academic departments and observatories and official communications of the IAU and ADASS. The listserv was one of a bundle of FITS services introduced in the late 1980s and early 1990s designed to increase the public presence of FITS on the internet: online software directory, ftp archive, and e-mail exploder. On May 30th 1991, Don Wells sent the initial message to the FITS usenet group (Figure 2).

Figure 2. Snippet of the first alt.sci.astro.fits message.30

In a continuation of the collegial academic style of peer committees and BoF groups, Wells demurred discussing FITS further until his colleagues Preben Grøsbol and Barry Schlesinger had weighed in on usenet. But the internet waits for nobody and it wasn’t long before ”_everyone_everywhere_ “ joined the news group. In contrast to formal communication style of academic governance, the news group sometimes brought hot “flame wars” into the discussion over FITS’ future31 (Figure 3).

Figure 3. Bit of ribbing on the listserv.32

In addition to being a departure from the formal style of academic governance, the comments underline the kind of expertise and tacit

April-June 2020

knowledge33 assumed to be shared within the confines of academic governance. Anyone with even a passing familiarity with astronomical imaging would know intuitively why GIF and TIFF are inadequate for transmitting astronomical data. Hence, an unintended consequence of the news group was to open a window for all manner of comments, many of which had to be addressed and rebutted. Despite the novel inquiry about GIFs and TIFFs addressed to “the internet,” the listserv was comparatively free of the “flame wars” that plagued many listservs and usenet groups. Most of the disagreements were about technical specifications, such as which keyword headers should be standardized or suggestions for implementing a coordinate system within FITS. In the following sections, we will turn to discussing the technical limitations of FITS, which would prove to be more consequential than inquiries about GIFs and TIFFs.

PROBLEMS LOCATING THE WORLD COORDINATE SYSTEM Good form in FITS always extends to include the human as well as the software readers.34

By 1998, FITS was well-entrenched as the default file format of astronomy. The email exploder was working as Wells intended, with various discussions and threads explaining what FITS is, how it can be used, and proposing changes in the future BoF sessions. But in 1998, a lingering and unresolved topic necessitated action: The World Coordinate System (henceforth WCS). In the WCS controversy, the semantic elements necessary to interpreting data began to cut against FITS’ syntactic function as an interchange format. World Coordinates serve to situate an object in a space, so that its location can be fixed. The WCS is a set of transformations that map locations in the sky to pixel locations in an image, or the reverse, from pixel images to positions in the sphere of the sky. These locations, or measurements at locations, can be multidimensional in form.35 The WCS can also be used to define wavelength transformations, for example for spectroscopy, the study of electromagnetic radiation emitting from stars.

47

Article

The original FITS specifications from 1981 described a method for locating the coordinates of image pixels but did not specify conventions for how to locate those coordinates in the sphere of the sky. This was a deliberate design decision that left the astronomer to decide how to map the image to the sky. Over time, differing groups of astronomers (optical, radio, X-ray, infrared, ground-based and space-based) had developed de facto, and incompatible, conventions for representing the sphere of the sky, cutting against FITS mandate as an interchange format. To solve some of these issues caused by the lack of definition in the FITS standards, some conventions were created by members of the FITS community that were nonstandards yet widely adopted, such as ad hoc versions of the WCS. Software libraries popped up to help with these coordinate transformations needed to interpret FITS images and fill in gaps left by the official FITS standard.36 Wells’ solution to the WCS problem was to create a listserv called wcsfits, to build consensus for a standardized approach in FITS using the WCS by including input from all observing astronomers in the proposals for addendums to FITS. The wcsfits listserv revealed a chaos of competing opinions about how a standardized WCS might be implemented. Implementing any coordinate system is not a trivial undertaking as the difficulty lies in the nature of locating an image in the sky, not in representing it in the format. Vexing issues arose as to how the tools could read the WCS components in the FITS files: should FITS ignore some methods and read others? How much complexity should be built into FITS headers? The disagreements lie in both the organizational structure of the file and the differing ways groups of astronomers had organized their measurements and orientations towards the sky. Throughout the late 1990s, various approaches were discussed at length on the wcsfits listserv, without a clear consensus. Yet by 2002, an initial attempt at a standard approach had been made. Two papers were published in 2002, one in 2006, and another in 2015.37 These new WCS conventions provided a standard method to map physical coordinates in the sky. Though the WCS issues played out for

48

years (and still do) within FITS governance, and certain functionality has been added, for most it still requires effort on the part of the researcher to make it work. Griesen38 argues that the failure to adopt a WCS system in FITS is due to lacking the will to settle on something. FITS critics seem to think that the failure to adopt WCS guidelines is another argument against FITS. WCS exposes the differences in subfield work and pushes FITS to the existential limit of what FITS should be- should FITS be a Swiss army knife or provide directions to build one? It was too late, though, unlike the way FITS was developed (by a handful of people), using the listserv to create consensus for a standardized way to represent the WCS was a failure. By the time a standard was proposed, subfield-specific methods of implementing the WCS were entrenched and many in astronomy had moved on to new forms of astronomical analysis.

ANALYTIC PIPELINES AND THE RETURN OF THE INTERCHANGE PROBLEM “Once FITS, always FITS” doctrine, which has been utilized to effectively freeze the format, was a mistake in our opinion. Adherence to the doctrine, and lack of any means to version the format in a machinereadable manner, has stifled necessary change of FITS.39

As FITS transitioned from an interchange format capable of bridging subfields and computing systems into the de facto archiving standard for astronomical data, new astronomical tools, such as Infrared Reduction and Analysis Facility (henceforth IRAF),40 were built on top of FITS. Meanwhile, tools imported from outside astronomy, such as relational databases that served FITS files, and object-oriented programming languages that took FITS as objects, became commonplace within astronomy. These changes were driven by what has come to be called dataintensive science, which has seen the velocity and volume of data increase by several orders of magnitude since 1981 when astronomers went up the mountain to observe and came down the mountain with a box full of magnetic tapes to analyze. Today, astronomers are more likely to divert a stream of data from a server.

IEEE Annals of the History of Computing

Consequently, many astronomical research groups now employ statisticians, data scientists, and archivists simply to stay afloat on the everincreasing flow of data in astronomy.41 And rather than being a question of whether or not the magnetic tapes one brings off the mountain are compatible with one’s institution’s computers, analysis takes the form of complex pipelines in which data is transformed and, in astronomy parlance, reduced in a series of steps. The WCS controversy illustrated why FITS is not an ideal format for analytical work in astronomy today. Several recent papers have outlined issues in current computational environments in which FITS does not perform well.39 For example, if one has a FITS file of 100 GB, which is larger than the RAM memory in the computer, the file will be read slowly. To quickly read large files, a modern container is required to handle a large array of data. Astronomers introducing new kinds of analysis wanted more analytical options,42 in particular, they chafed against the FITS structure that determines when analysis takes place.38 As another astronomer noted, FITS works elegantly with the volume and velocity of data common in the 1980s and 1990s, but warned that “as data sets get bigger it is going to start failing.”43 Data sets have gotten bigger and the increasing volume and velocity of data within astronomy has seen the interchange problem return in a surprising way–as a new chaos of programming languages and file formats deployed in astronomical data pipelines. A data pipeline is a set of actions to designed to extract and transform data for analytical use.44 The interchange problem increases as data moves further down the pipeline, where the the needs and inclinations unique to a research group accumulate in the bits and pieces of programming languages and formats a pipeline is constructed from. FITS is still omnipresent, but today often serves as a point of departure, not an endpoint. Three file formats in particular, each using FITS as a touchstone, are in use within astronomy: the Hierarchical Data Format (henceforth HDF5) stemming from federally funded organization attempts at creating a universal data analysis format, VOTable, an attempt at building a broad set of XML-based tools for astronomical archiving and

April-June 2020

web-based analysis, and Advanced Scientific Data Format (henceforth ASDF) a newer file format with many similarities to FITS, such as a human readable headers with binary structures for data. HDF5 is used throughout many scientific fields for high speed computing. HDF5 emerged out of a score of file formats proposed for use in the Earth Observing System project headed by NASA45 and over time matured into a format able to store different kinds of information. It can handle large files in the terabyte range and is therefore favored by many data scientists. VOTable format was designed for astronomers to continue the work of the Virtual Observatory methods of making online data interoperable.46 VOTable uses XML as a standard to represent data as a set of tables, with XML being the interoperability piece. However, it has a major drawback, in that VOTable has no method to handle binary data. While HDF5 addressed the needs of data scientists and VOTable the needs of taking astronomy online for engaging citizens in astronomical analysis,47 ASDF was developed as a direct replacement for FITS. Publicly announced in 201548, like FITS, ASDF was created by astronomers49. ASDF resembles FITS in structure and function, containing a human and machine-readable header, rendered in YAML and is backward compatible with itself for archival purposes. Unlike the FORTRANderived FITS headers, however, YAML headers support hierarchical information, thus overcoming FITS’ limitations in working with WCS coordinates. Unlike during the 1970s, today computers are highly interoperable. Yet within astronomical pipelines, software interdependencies, software libraries relying on other software libraries for basic functions, have made for extremely complex layers of code that are brittle and prone to breakage.50 One widely used strategy to overcome this problem is organizing software infrastructure via an open source Github repository, where interdependency problems can be crowdsourced. Like FITS, ASDF is reliant on a tremendous amount of volunteer work. On the ASDF Github repository, volunteers are welcomed in name of openness in astronomy and encouraged to make contributions. The Github review mechanism sets up a system of governance, where contributors to the project are welcomed, but

49

Article

their contributions might be rejected, ignored, or folded into the project by a closed circle of core developers. In a sign of emerging institutional acceptance, NASA’s delayed but hopefully soon to be launched James Webb Space Telescope will offer data in both ASDF and FITS formats. Since the focus of money spent on astronomy is geared towards the instrumentation, this institutional support augers well for ASDF’s long-term future, and points toward the possibility of institutionally based governance in ASDF’s future.

FUTURE(S) OF ASTRONOMICAL INFRASTRUCTURE? Code is hard, code rots. So if your data depends on having code available like this opaque binary file, that’s bad. And that’s where astronomy. . . It’s FITS files. Well, not everything needs to be in a FITS file. It certainly lets you package up both data so that they will be around for a long time. And I hear this in other fields like “we don’t have anything like that, how do you guys do it?” Physics, we don’t have anything like that. Biology, we don’t have anything like that. It’s all in people’s notebooks whatever.– research scientist at a U.S. research lab51

FITS is a simple format consisting of an ASCII header containing metadata, along with optional extra units,52 and data units consisting of images and tables of ASCII or binary data. FITS has managed to serve as an interchange format between machines, subfields, and international differences. Though informal and rickety (agonizingly slow for some, overly stable for others), FITS remains the de facto standard for archival and interchange in astronomy today.53 It is estimated that there are more than 1 billion FITS files in various archives around the world.42 Like the format itself, which is flexible enough to accommodate radio astronomers, optical astronomers, and Vatican archivists, and durable enough to ensure the original FITS files produced forty years ago are still usable. Likewise, the infrastructure of FITS governance, as complex and the format is simple, has proven flexible and durable enough to see FITS through sea changes in the nature of computation and astronomical research. The process for amending and adding to the standard assures broad community

50

participation, and although this sometimes makes the process of change rather slow it helps to assure community support and compliance. Today, FITS crawls along as it has for decades, working behind the scenes on every astronomer’s desktop, every telescope’s archive, and in every astronomical database, carefully maintined and updated by a dedicated and tight-knit group of aging astronomers. FITS began its life as an interchange and subsequently became an archive format, but inevitably it also became a format for analysis. As a format for analysis, FITS is tied to a particular computing paradigm, imperative programming, and the computer language, FORTRAN, that were the common currency of astronomy in the 1970s. By 21st century standards, FITS is antiquated, difficult to maintain, and tied to computing’s past, not its future. Astronomer’s in graduate programs today learn Python, not FORTRAN. Further, FITS is tied to a governance infrastructure that was the common currency in the academia of the late 1970s, but seems increasingly antiquated given changes to both the discipline of astronomy and working conditions in academia. Yet paradoxically, despite its technical obsolescence, in the era of data-intensive science, FITS has become increasingly useful for mining the burgeoning astronimcal archives and serving as a point of departure for emerging forms of astronomical analysis. What does it mean for FITS governance that FITS is both increasingly useful and increasingly antiquated? Infrastructure for scientific disciplines expected to span generations requires care and attention to both the technical details and governance infrastructure. It must be attentive to, and change, with the context of the academy and specific disciplines. Can a cohort similar to the original FITS cohort be created, and if so, what form would it take? Is it possible to replicate the kind of stability long academic careers and stable academic associations gave to FITS through ad hoc projects organized over the internet? No single organization is in charge of thinking through next generation astronomy standards. Today, the precarious postdoc and research scientists who do the everyday work of scientific computation lack the time, and increasingly the incentive, to volunteer for

IEEE Annals of the History of Computing

academic governance work, such as improving and maintaining FITS, in a profession they may leave after a few years. The biggest threat to FITS governance and, therefore, its future in astronomy, might be the upcoming generational turnover in the committees and working groups that care for, maintain, and update FITS. Despite the uncertainty of FITS future care and governance and the emergence of new forms of astronomical analysis, the observer’s wisdom embodied in the edict “once FITS, always FITS” remains as relevant as ever to astronomy. The complaints and handwringing over FITS obsolescence and inflexibility, astronomers who venture into other fields often find themselves mired in a chaos of competing proprietary formats that mirrors the situation in astronomy during the 1970s. As one astronomer replied when asked what astronomy would be without FITS, “it would be like materials science is today. . . if there are ten different vendors’ machines down the hallway here, likely ten different formats are coming out of them, not interoperable, not with common metadata standards.”54 Perhaps it is not surprising, then, that the newest file format in astronomical computing today takes inspiration from the oldest. ASDF, created by astronomers for astronomy, can be understood as a technical update of FITS set on a n ew governance infrastructure with a new collaboration and governance style more in keeping with the norms of open source software communities. Despite ASDF’s intention to be a modern update of FITS, it should not be viewed as a replacement for FITS. Billions of FITS files are stored on servers across the world. As astronomy archives’ usage increases, evidenced by citations in publications, FITS files themselves are increasingly irreplaceable.

of the Galactic Center Group at UCLA for additional astronomy research.

& REFERENCES 1. A.

Astronomer,

private

communication

with

B. Boscoe, 2018. In writing this article, we have drawn on three sources. First, the publically available archival material housed at NRAO (https:// www.cv.nrao.edu/fits/). Second, the corpus of research on data-intensive science, which includes interviews with more than one hundred working astronomers, generated over the last decade by the Center for Knowledge Infrastructures at UCLA. Third, interviews with more than forty working astronomers, ranging from graduate students to emeritus, focused on the place of FITS in their current work conducted by Boscoe from 2015-2019, 2004. 2. S. Allegrezza, “Flexible image transport system: A new standard file format for longterm preservation projects?,” Vatican Library, Vatican City, 05-Jul-2012. Vatican archivists searched high and low for the perfect format to preserve their precious collection of images of texts and other articles for eternity. They found their answer in another group also looking at the heavens. See also N. Barbuti, “Considerations on the preservation of base digital data of cultural resources,” in Digital Libraries and Archives. vol. 354, M. Agosti, F. Esposito, S. Ferilli, and N. Ferro, Eds., Berlin, Germany: Springer, 2013, pp. 13–20, and C. Barbieriet al., “Digitization and scientific exploitation of the italian and vatican astronomical plate archives,” Exp. Astron., vol. 15, no. 1, pp. 29–43, 2003. 3. D. C. Wells, E. W. Greisen, and R. H. Harten, “FITS: A flexible image transport system,” Astron. Astrophys. Suppl. , vol. 44, pp. 363–370, 1981. From now on in this article, when we refer to the

ACKNOWLEDGMENTS This research work was funded by the Alfred P. Sloan Foundation, Award# 2015-14001: If Data Sharing is the answer, what is the question? We thank Christine Borgman, Milena Golshan, Morgan Wofford, Peter Darch, and Cheryl Thompson of the UCLA Center for Knowledge Infrastructures. We also thank and Tuan Do, Mark Morris, Abhimat Gautam and all members

April-June 2020

seminal 1981 paper, we refer to this first paper defining FITS. 4. M. D. Gordin, Scientific Babel: How Science Was Done Before and After Global English. Chicago, IL, USA: Univ. Chicago Press, 2015. Also of interest, see P. Grøsbol, R. H. Harten, E. W. Greisen, and D. C. Wells, “Generalized extensions and blocking factors for FITS,” Astron. Astrophys. Suppl., vol. 73, no. 3, pp. 359–364, Jun. 1988.

51

Article

5. “FITS standard page,” FITS Standard Document: The official reference document that defines the

infrastructure: Design and access for large information

requirements for FITS format data files, Dec. 2017.

spaces,” Inf. Syst. Res., vol. 7, no. 1, pp. 111–134, Mar.

Accessed: Jun. 29, 2018. [Online]. Available: https://

1996. Star and Ruhleder famously identify eight

fits.gsfc.nasa.gov/fits_standard.html.

dimensions of infrastructure. Despite giving the caveat

6. M. W. Wiseet al., “Data and documentation,”

that these dimensions are to treated relationally as

Transactions IAU, vol. XXXA, Reports on

“whens” rather than as fixed “whats,” their dimesions are

Astronomy 2015-2018. As we will see later in the

often taken up as universal tenents of infrastructure. This

paper, astronomers are now considering

misreading elides the role of the most mundane

alternatives to the aging FITS, and from the BoF

infrastrucural elements, such as FITS, in making

formed DRWG to expand from just FITS and to

infrastructure resilient. See, for example: C. L. Borgman,

“to manage a careful and minimally disruptive

A. E. Sands, P. T. Darch, and M. S. Golshan,

transition from FITS to more modern and capable

“The durability and fragility of knowledge infrastructures:

data representations.” 2018. [Online]. Available:

Lessons learned from astronomy,” Proc. Assoc. Inf. Sci.

https://www.iau.org/static/science/scientific_bodies/ commissions/b2/commission-b2-triennial-report-

Technol., vol. 53, no. 1, pp. 1–10, 2016. 12. G. C. Bowker, “Second nature once removed: Time,

2015-2018.pdf

space and representations,” Time Soc., vol. 4, no. 1,

7. “IAU FITS working group activities,” 2014. Accessed:

pp. 47–66, Feb. 1995.

Apr. 24, 2018. [Online]. Available: https://fits.gsfc.

13. J. Law and A. Mol, “Globalisation in practice: On the

nasa.gov/iaufwg/iaufwg_activity.html. FITS usenet

politics of boiling pigswill,” Geoforum, vol. 39, no. 1,

newsgroup was the first internet-based communica-

pp. 133–143, Jan. 2008.

tion tool used for public outreach. An email exploder

14. B. Latour, Pandora’s Hope: Essays on the Reality of

followed enabling people to use email to communi-

Science Studies. Cambridge, MA, USA: Harvard Univ.

cate, and then listservs. At present, the listserv

Press, 1999. 15. J. Sterne, MP3: The Meaning of a Format. Durham, NC,

remains as a means to share messages with FITS

USA: Duke Univ. Press, 2012.

followers. 8. A. M. Kjaer, Governance. 1st ed. Malden, MA, USA:

16. A. L. Russell, Open Standards and the Digital Age. Cambridge, U.K.: Cambridge Univ. Press, 2014.

Polity, 2004. Also see J. V. Baldridge, “Models of

17. B. Astronomer, Private communication with B. Boscoe,

university governance: bureaucratic, collegial, and

M. Scroggins, 2018.

political,” Sep. 1971. 9. C. M. Kelty, Two Bits: The Cultural Significance of

18. D. Cohen, “On holy wars and a plea for peace,”

Free Software. Durham, NC, USA: Duke Univ.

Computer, vol. 14, no. 10, pp. 48–54, Oct. 1981, doi:

Press, 2008. Open source writings include, inter

10.1109/C-M.1981.220208. 19. W. P. McCray, “The biggest data of all: Making and

alia, A. Fish, L. F. R. Murillo, L. Nguyen, A. Panofsky, and C. M. Kelty, “Birds of the internet,”

sharing a digital universe,” Osiris, vol. 32, no. 1,

J. Cult. Econ., vol. 4, no. 2, pp. 157–187, May

pp. 243–263, Sep. 2017. We encourage reading

2011, and S. K. Shah, “Motivation, governance,

P. McCray’s history of astronomy work. 20. C. Astronomer, Private communication with B. Boscoe,

and the viability of hybrid forms in open source

2018.

software development,” Manage. Sci., vol. 52,

21. D.

no. 7, pp. 1000–1014, Jul. 2006. 10. B.

52

11. S. L. Star and K. Ruhleder, “Steps toward an ecology of

Larkin,

“The

politics

and

poetics

of

Dickinson,

“Remembering

the

‘World

War

I Eclipse,’” Universe Today,. The outbreak of the Great

infrastructure,” Annu. Rev. Anthropol., vol. 42,

War had coincided with a rare solar eclipse in

no. 1, pp. 327–343, Oct. 2013. See also P. Harvey

continental Europe and the much of the observational

and H. Knox, Roads: An Anthropology of

record was lost to the war. The IAU was international,

Infrastructure and Expertise. 1st ed., Ithaca ;

to a point. Astronomers from the Central Powers were

 zek, London, U.K.: Cornell Univ. Press, 2015. R. Mra

intentionally omitted from joining. Germany, for

Engineers of Happy Land: Technology and

example, a leading astronomical power in the 19th

Nationalism in a Colony. Princeton, NJ, USA:

century, was banned from the IAU until after the

Princeton Univer. Press, 2002.

Second World War, Aug. 2014.

IEEE Annals of the History of Computing

22. R. M. West, Proceedings of the Eighteenth General Assembly: Patras 1982. Berlin, Germany: Springer, 2012.

33. H. M. Collins, “The TEA Set: Tacit knowledge and scientific networks,” Sci. Stud., vol. 4, no. 2, pp. 165–185, Apr. 1974.

23. P. Gro/sbol, R. H. Harten, E. W. Greisen, and D. C. Wells,

34. [Online]. Available: https://listmgr.nrao.edu/pipermail/

“Generalized Extensions and Blocking Factors for FITS,”

fitsbits/, 1998. The Vatican archive announced in

Astron. Astrophys. Suppl., vol. 73, no. 3, pp. 359–364,

the April 2018 newsletter their intention to archive

Jun. 1988.

the Vatican’s FITS files for 1,000 years on long-last-

24. W. P. McCray, Giant Telescopes: Astronomical

ing archival film format in the Artic World Archives.

Ambition and the Promise of Technology. Cambridge,

The announcement spurred a conversation on the

MA, USA: Harvard Univ. Press, 2004.

FITS listerv lamenting the lack of funding for such

25. “FITS resources,” 2018. Accessed: Aug. 18, 2018.

initiatives in astronomy, for this message see

[Online].Available: https://fits.gsfc.nasa.gov/

[Online]. Available: https://listmgr.nrao.edu/

fits_resources.html. The FITS leaders were also

pipermail/fitsbits/2018-April/003030.html, and for

excellent documentarians. Websites of links to

the Vatican announcement, see [Online]. Available:

papers, memos, meetings, listserv archives and all

www.vaticanlibrary.va/newsletter/20183EN.pdf,

manner of information concerning FITS are still available on the internet today, this link is a great place to start.

scroll down to page 9. 35. R. M. Green, Spherical astronomy. Cambridge [Cambridgeshire]; New York: Cambridge University

26. C. L. Borgman, “From acting locally to thinking globally: A brief history of library automation,” Library Q., Inf., Commun., Policy, vol. 67, no. 3, pp. 215–249, 1997. Even a purely syntactical format such as MARC records, an ISO standard based

Press, 1985. 36. “Registered FITS Conventions,” 2012. Accessed: Jul. 5, 2018. [Online]. Available: https://fits.gsfc.nasa.gov/ registry/tpvwcs.html. 37. “FITS WCS page,” Apr. 2018. Accessed: Jul. 5, 2018.

format used to describe items in libraries, is not

[Online]. Available: https://fits.gsfc.nasa.gov/fits_wcs.

immune to the problem of unspecified a priori

html.

parameters.

38. E. W. Greisen, “FITS: A remarkable achievement in

27. A. McKenzie, “INWG and the conception of the

information exchange,” in Information Handling in

internet: An eyewitness account,” IEEE Ann. Hist.

Astronomy–Historical Vistas, A. Heck, Ed.

Comput., vol. 33, no. 1, pp. 66–71, Jan. 2011.

Dordrecht, the Netherlands: Springer, 2003,

28. C.

Barbieri

et

al.,

“Digitization

and

scientific

exploitation of the italian and vatican astronomical

pp. 71–87. 39. B. Thomaset al., “Learning from FITS: Limitations in

plate archives,” Exp. Astron., vol. 15, no. 1, pp. 29–43,

use in modern astronomical research,” Astron.

2003, doi: 10.1023/B:EXPA.0000028168.26442.30.

Comput., vol. 12, pp. 133–145, Sep. 2015.

29. H. M. Collins and R. Evans, Rethinking Expertise. Chicago, IL, USA: Univ. Chicago Press, 2007. 30. 1991. [Online]. Available: fits.gsfc.nasa.gov/fitsbits/

40. D. Tody, “IRAF in the Nineties,” Astron. Data Anal. Softw. Syst. II, vol. 52, 1993, Art. no. 173. 41. S. Murray, “The LSST and big data science,”

saf.91/saf.9105. We laud the astronomers for keeping

Astronomy.com, Dec. 2017. Accessed: Aug. 19. 2018.

and archiving their communications!

[Online]. Available: http://www.astronomy.com/news/

31. A. Hocquet and F. Wieber, “Mailing list archives as useful primary sources for historians: Looking for flame wars,” Internet Hist., vol. 2, no. 1/2, pp. 38–54, Apr. 2018. According to Hocquet and Wieber, the listserv is valuable to the historian, giving a corpus “constituted

2017/12/the-lsst-and-big-data-science. 42. J. Mink et al., “The past, present and future of astronomical data formats,” Nov. 2014, arXiv:1411.0996. 43. D. Astronomer, Private communication with B. Boscoe, 2018.

of a sort of middle layer of discourse between orality

44. D. M. Ritchie, “The UNIX system: The evolution of the

and formality” [p. 39] where the standards of neither

UNIX time-sharing system,” AT&T Bell Laboratories

orality nor formality had complete hold and

Technical Journal, vol. 63, no. 8, pp. 1577–1593, 1984,

misunderstanding was always near. 32. W. Pence, 2018. [Online]. Available: https://listmgr. nrao.edu/pipermail/fitsbits/2018-April/003030.html.

April-June 2020

doi: 10.1002/j.1538-7305.1984.tb00054.x. 45. “The HDF group,” 2006. [Online]. Available: www. hdfgroup.org.

53

Article

46. F. Ochsenbein et al., “IVOA recommendation– VOTable format definition,” Sep. 2013. For an explanation of Virtual Observatory work, see [Online]. Available: www.ivoa.net. 47. M. J. Raddick et al., “Galaxy zoo: Exploring the

53. R. J. Hanisch et al., “Definition of the flexible image transport system (FITS),” Astron. Astrophys., vol. 376, no. 1, pp. 359–380, Sep. 2001. 54. E. Astronomer, Private communication with B. Boscoe, 2018.

motivations of citizen science volunteers,” Astron. Educ. Rev., vol. 9, no. 1, 2010, Art. no. 010103. 48. P. Greenfield, M. Droettboom, and E. Bray, “ASDF: A new data format for astronomy,” Astron. Comput., vol. 12, pp. 240–251, Sep. 2015. 49. “ASDF–Advanced Scientific Data Format–asdf v2.1.0. dev1447,” 2020. Accessed: Aug. 19, 2018. [Online].

Michael Scroggins is currently a postdoctoral scholar with the Center for Knowledge Infrastructures, University of California Los Angeles, CA, USA. He graduated from Columbia University with a Ph.D. degree in Cultural Anthropology. Contact him at [email protected].

Available: https://asdf.readthedocs.io/en/latest/. 50. D. Muna et al., “The astropy problem,” arXiv:1610.03159, 2016. Accessed: May 2, 2020. [Online]. Available: http:// arxiv.org/abs/1610.03159 51. Interview in Center for Knowledge Infrastructure archive, by S.K., 2011. 52. B. Schlesinger, Draft FITS section for formats

Bernadette M. Boscoe is a postdoctoral researcher with the School of Information, University of Washington, USA, where she is also a data science postdoctoral fellow with the eScience Institute. She is the corresponding author of this article. Contact her at [email protected].

comparison booklet, May 1992.

54

IEEE Annals of the History of Computing

Department: Anecdotes

Queens of Code Eileen Buckholtz Director, Queens of Code Project

INTRODUCTION TO THE QUEENS OF CODE

& QUEENS OF CODE is a women’s technology history project—a collection of stories, experiences, and insights from women who worked in information technology at the National Security Agency (NSA) in the 1960s, 1970s, and 1980s. NSA’s computing women programmed and managed the most sophisticated systems of their day and I was one of them. I started this project in 2018 to collect the stories of the agency’s women technology pioneers and recognize their contributions because I believed that if we did not document these stories now while many of us are still living, our history would never be told. The National Cryptologic Museum and NSA’s historians offered encouragement. I reached out to women I had worked with, and dozens signed up. Participants were asked to complete a detailed questionnaire and write their stories. All material had to be approved through NSA’s prepublication review. We have been networking online for almost two years and have more than 75 women in the group. The goals for the project are recognition of the Queens of Code in the history of computing, expanding the understanding of how women worked in early computing, and inspiring more young women to pursue STEM careers. We are sharing our stories in presentations, articles, and interviews. Because these NSA women’s jobs were often top secret and they worked on the most sensitive national security programs, they could not Digital Object Identifier 10.1109/MAHC.2020.2982751 Date of current version 29 May 2020.

April-June 2020

discuss what they did, even with their families. In many cases, they could not even confirm they worked for NSA. They and their computing activities have been, for practical purposes, a secret for more than 50 years. Women have always been in the workforce— although their contributions to science have often gone unrecognized. In the 20th century, women worked for the U.S. government and military, not just in clerical, nursing, and other “women’s” positions, but in specialized technical fields such as cryptology, mathematics, and computing. The U.S. military during World War II actively recruited educated and talented women, including those from some of the best colleges, to fill critical vacancies and to “free a man to fight.” These women often found themselves doing tedious work, but gained a foothold in the technical workplace. According to Liza Mundy’s Code Girls, over 10,000 women were a critical part of the cryptologic mission, some working with the early computing machines.1 In the U.S. many women who had technical skills were sent home after the war to free the jobs for men returning from war. More generally, women’s place in computer history has not been publicized because it has largely been HIStory, focusing on hardware and the male inventors,2 as I saw on my visit to the Computer History Museum in Mountain View, California, in July 2018. Fortunately, modern cryptology, in particular, was welcoming to women from the start. Elizebeth Friedman and Agnes Driscoll led the way in the 1920s and 1930s.3 The work of the “Code Girls” during World War II was critical for

Published by the IEEE Computer Society

1058-6180 ß 2020 IEEE

55

Anecdotes

winning the war. Like their contemporaries at NASA, whose story was told in the bestselling book and hit movie Hidden Figures, the women at NSA, walking in the footsteps of their World War II sisters, have broken ground from the 1960s on as they contributed to advances in computing in the world of cryptology.4 Many of the Queens of Code were recruited by NSA right after college and worked in computing technology for 30, 40, and even 50-year careers. I was one of those queens, hired in 1970, with one of the first undergraduate degrees in computer science in the country. Starting from data systems interns and rising to senior leaders and computer science experts, we were on the forefront of computer technology development. In the 1960s, 1970s, and 1980s, our agency had the most sophisticated computers in the world as well as the most challenging information processing requirements. By 1968, NSA had more than 100 computers spread over five acres of computer rooms.5 The inventory grew rapidly over the next decades as we and our male colleagues worked with many vendors to drive new system development to meet our big data processing needs. Our stories may also provide some insight to companies today that struggle to recruit and retain women in tech. In contrast to corporations and institutions in various other sectors, NSA did a lot right over a 50-to-60-year period to recruit, develop, and retain their computing women. They had learned from previous experience with the Code Girls during World War II that women were a valuable asset to their mission. They invested in us through training, intern programs, and advanced degrees, paid equal starting salaries for men and women, gave women responsibility and credit, promoted many women to senior management and technical positions, and provided a good work/life balance. Fortunately for us, most of the men we worked with were supportive as well. Of course, there were some struggles along the way, including a classaction lawsuit over fair promotion in the 1970s,6 but we prevailed. The Queens of Code made a daring leap into a new career field of computer science and found innovative, exciting, and rewarding careers that contributed to the hightech world we live in today The rest of this article highlights some of the experiences I, and many other women, shared working on our first computers.

56

FIRST ENCOUNTERS OF A BINARY KIND If you grew up in the 1980s or 1990s as part of the millennial generation, your first experience on a computer might have been with a personal computer at home or at school. You might have learned to program in basic using my Micro Adventure books7 on an Apple II, Radio Shack TRS-80, Atari, or IBM PC. If you are part of Generation Z, you probably played games on your first computer tablet or smart phone maybe as early as a toddler. E-books, apps, and online shopping and learning are things you took for granted. That was not the case when the Queens of Code were young. The ARPANET (the early version of the Internet that had just begun to come online in 1969) only connected some dozens of government agencies, universities, and other research organizations, and the World Wide Web had not been invented. Back in the early 1970s, one of our offices did have a terminal that we could use a modem to dial into the National Bureau of Standards’ ARPANET. From NBS, we could connect to the Stanford Research Institute—and it took dozens of steps to send a line of text along with manually calculated checksums (a digit that was the sum of the other digits in a piece of data used to detect errors). When we were growing up, there was little digital computing technology in the schools we attended before college. Pocket-size calculators made their debut in the 1970s. Before then, in high school or college, we used a slide rule (the manual device invented in 1620) for math, chemistry, or physics courses. Many of us were 18-to-22-year olds when we met our first computer, perhaps an IBM 1620, 1401, or even 360 (after its release in 1964) at our college or university. Often the Queens of Code’s first computing experiences were on their initial assignment at NSA or at college. These computer installations could be huge and expensive, especially those in NSA’s extensive basement sometimes taking up spaces as big as a couple of basketball courts, cooled by water under the floors to make the rooms so cold that you had to wear a heavy sweater or jacket when working there. Some of our first computing experiences were on computers with limited capacity and programming done in assembly language or even octal, and that was not easy. We had to be crafty to

IEEE Annals of the History of Computing

make the programs work within the constraints. FORTRAN, the first commercially available computer language to use a compiler was released in 1957. A compiler meant that the code could be written with higher level and easier to manage commands that would automatically generate the assembly language or machine code needed for a specific machine. FORTRAN was designed to provide a language for the scientific community, and NSA certainly fit in that box. As computer technology advanced and memory size increased and became less expensive, programmers could write code with less computer specific restrictions. Dottie Blum, a legendary computing woman at our agency, was using FORTRAN as early as 1954 even as it was being developed by John Backus and team at IBM. At first, people wondered if using a compiler would produce code as efficient as writing in assembly language. But over time, computer speed and memory size increased and convenience won out. Another benefit was that programs could be ported (moved over) to run on other machines that had a FORTRAN compiler. It was a big improvement over having to rewrite programs in another assembly language every time a new computer came into our collection.8 Programming was a little like cooking. You had input (like your ingredients) and then steps that processed the ingredients. If all worked well, you have something to eat for supper. Fortunately, I was a better programmer than a cook. Our first programs were either assignments at school or “toy” programs we were assigned to learn how the computer worked. On the earliest computers from the 1950s like the special purpose ones built in house, there was only the basic documentation, so you had to figure things out for yourself. Sarah, one of our first programmers, used to say that when she started at NSA the computers took up a whole room and you were lucky to find a small notebook with instructions. By the time she retired, the computers were small enough to fit on your desk, and you had a bookcase of manuals and online documentation. In our environment, the programming process worked as follows. The first step was to define the problem. In our case for application programs, this meant talking to the analyst to understand the problem that needed solving. The problem was often to automate a timeintensive manual process such as an attack on a cryptographic code we had collected by

April-June 2020

analyzing signals or language translation. NSA processed tons of data to produce intelligence reports for government decision makers including the President and the military. NSA was doing “big data,” long before the phrase was coined in 2005. Programs were written to support requirements at the time. The programmer would then break the process down into small steps that would provide a solution. Programmers often use flowcharts to block out the steps that need to be taken. We used plastic templates back then to draw the flow charts.9 Now there are many software tools and applications to help with program design. Next, we had write the code in a programming language like FORTRAN, PL1, or C or in assembly language in the earlier days. Then, we had to debug it, resolving all the problems that we could find. After that, we tested with our realworld users; and, when all was working, officially declared the program live. Of course, there would always be more bugs that popped up, and we had to fix those in a timely manner. At the agency, system programmers who worked on the operating systems and networking were in the C (for Computer) Organization (which was later reorganized and renamed T (for Telecommunications and Computer) Organization. It seemed that every three or four years we have a major reorganization, sometimes corresponding to a new Director’s arrival. Some application programmers started out as part of C, but later moved out to sit with the users in the production organizations. All the reorganizations and reassignments were confusing. One of my bosses had a sign in his office that read, “Perfect reorganization is only achieved by groups on the verge of collapse.”

LEGACY QUEEN’S FIRST ENCOUNTER WITH A COMPUTER Dottie Toplitzky Blum, 1950 Dottie had worked with the Electronic Adding Machines (EAM) equipment and the Army’s version of the BOMBE, an electromechanical device developed by Joe Desch of National Cash Register during WWII to decode Enigma messages. Another of Dottie Blum’s earliest binary encounters was with the Standards Eastern Automatic Computer (SEAC), which was built in Washington, DC, USA, for the National

57

Anecdotes

Bureau of Standards. The SEAC was one of the first U.S. stored program computers. Dottie then worked for AFSA, the Armed Forces Security Agency, NSA’s predecessor. AFSA did not have their own computer but the support organization did manage a number of calculating and cipher machines including the Navy’s and Army’s BOMBEs that were used in the war effort. These earlier devices were not actually computers since they lacked memory or ability to do anything outside of their limited computational functions such as compiling and comparing text, searching for cribs (plain text), or calculating statistics. It was 1950 and Dottie and Sam Snyder, one of her coworkers whose computer history writings documented this story, got an urgent request from the Navy’s Communications Security Division. The job required the verification of a few hundred involutory 4 x 4 matrices10 that were used in the Navy callsign system. The SEAC’s memory was only 512 (45 bit) words, which was pretty limiting. Back then, they had to negotiate time on the SEAC to debug the program, and the only time NBS would allow them to purchase (at $24 an hour) was after midnight or on Sunday afternoons. The program was written but they needed test data. Dottie, who was working for the Machine Production Organization as an IBM specialist, produced thousands of random numbers on punched tape to be used to test the application. The SEAC took between 8 and 15 seconds to process each matrix and then printed out the ones that met the “useful” criteria. With a lot of work and some late nights and weekends, Dottie and Sam got the information to the Navy in a timely manner to help solve the problem. Sam said, “Those who participated in this task found the experience ‘frustrating, exhilarating sense of accomplishment and participation in making history.”11

MORE FIRST-HAND ENCOUNTERS FROM OUR QUEENS OF CODE Carol McWilliams, 1967–1970 My first programming experience was assembly language on a CP818 (UNIVAC 1224) for field installation. We “wrote” our programs on a Kleinschmidt—something like a typewriter, but it produced punched paper tape with one

58

instruction per line (e.g., “clear register”). You could fix an error by wrapping Scotch tape over the holes in the line and repunching the line! Fortunately, the readers were not sensitive to the opacity of the tape, just the holes. The resulting paper tape was wrapped butterfly style in a figure eight with a paper clip in the center and stored until you had time on the computer.12 The first programs I wrote were standalone processes. I had data input from magnetic tape, ran the program, and produced data output. I scheduled computer time and, when it was my time, I took my paper tape and mag data tape to the computer room. After loading the paper tape into memory, I put my magnetic tape on the spool and initiated the program. When I was done, I took my program tape, mag tape, and results off the computer, cleaned the heads for the next programmer, and took everything back to my desk to assess the results and debug my program. Very much a hands-on process!

Eileen Buckholtz, 1968-1969 I had transferred to Ohio State University (OSU) as a math major and was taking a fourth course in calculus while struggling with the theoretical proofs in the class. I remember the professor covering a big blackboard with the proof of the Heine–Borel Theorem. He got near the end, realized that he had made a mistake and started to erase half his scribblings. My eyes were glazing over. What was I doing here? Later that afternoon, I heard that OSU was opening their computer science department and they were looking for students. My boyfriend Howard was in engineering and he heard the same thing. Turns out they were offering degrees in both the Arts and Sciences Department, where I was enrolled as well as an engineering computer science degree. We both signed up, became OSU’s first computer science graduates, and have been computing together for over 50 years. There were only several dozen students in the first computer science classes. It was love at first byte for me when I took my first programming course. The initial assignment was a simple sort. The next was to use a random number generator to simulate shuffling a deck of cards and a matrix for holding all the hands. The idea that you could learn a language like FORTRAN and make an enormous computer do your bidding

IEEE Annals of the History of Computing

with structured commands was just fun. As we got into more advanced programs, it became challenging as well. An IBM 360 installation including CPU, tape drives, IO controllers, and other peripherals were housed in a big computer room that took up much of an upper floor of the engineering building. We could see into the computer room through big windows but we were not allowed to go in. After class, we would design our programs and then punch up Hollerith cards (one line of code per card) on keypunch machines, submit them over the counter, and then wait 5 or 6 hours for them to run and get our output back. If we made an error, we had to correct that and submit again. No wonder the four women in the computer science program were all dating guys also in the program. Who else would want to spend Friday and Saturday nights debugging programs?

Elaine Mills, 1965 As part of a work-study program at Towson State College (known today as Towson University), I was studying to become an elementary school teacher. I was privileged to be assigned to a special project to “computerize” all the records for the five Maryland state colleges. Using Hollerith cards and becoming proficient in writing FORTRAN programs in the mid-1960s was a “blast” that actually proved to be a tremendous personal boost a few years later at NSA.

Kathleen Jackson, 1967 My orientation started with a tour of the “basement,” a huge area where the computer I would be using, the UNIVAC 1108, was located. The system was so large that it nearly filled the whole computer room, since it had several printers and other pieces of peripheral equipment attached to it. My job was to remove computer printouts from one of the printers, review the data, look for data “anomalies,” and adjust the FORTRAN software as needed to fix them. It seemed challenging and interesting at first but, as the days and weeks rolled by, reviewing rows and rows of 1s and 0s became a little tedious to put it mildly. However, I persisted. After completing my tour, I looked forward to my next assignment. Over the years, I sometimes thought about that initial assignment, and how different

April-June 2020

it was from all the other work I had done at the Agency. Several years later, I came across a report that contained Agency historical information. It included information about the data that was processed on that UNIVAC 1108 computer I was supporting during my first assignment in 1967. This report identified how critical those data were to national security at that time. Suddenly, I became acutely aware that the many hours I had spent reviewing those 1’s and 0’s were definitely worth the challenge of the task. In the end, I determined that this work was probably some of the most significant work I did during my entire career at the Agency. I took pride in knowing that this work was very important to the security of our nation.

Kathleen Reading, 1982 “Oh great, another girl!” Imagine hearing those words upon meeting your supervisor for the first time. I was 21 years old and just beginning my 34-year career in the Information Assurance Directorate (IAD), in the Agency’s print shop. I was taken aback by my boss’s comment, but did not say anything as I was just starting a new job and did not know what to expect. I do remember thinking I was going to do everything in my power to change my boss’s mind about what “girls” could do. My job title was “Reproduction Worker,” and I was one of three women working in the shop. I found that job title pretty funny. I first started working in the bindery, and then also ran a printing press, large Xerox machines and printers, and eventually worked in the Electronic Printing and Publishing (EPP) branch. In the EPP branch, for the first time, documents to be printed were sent electronically via computer by Agency customers; and documents also were sent electronically to the printers for printing. One of the documents printed on the night shift was a daily report that was couriered each day to the White House. As it turns out, I did prove to my boss that women are good workers. I was promoted several times, and was also one of the first women selected to participate in the Agency’s first ever production trade program. Mary Clulow, 1977 “I will rule these machines; they will not rule me.” Quietly determined, I spoke these words

59

Anecdotes

late one evening at work while trying to complete a typing task. I was using an IBM Magnetic Card Selectric (aka MagCard) typewriter. This was quite a bold statement for an entry level Clerical Assistant at the National Security Agency. However, I had been challenged by this machine more than once since learning to use it two years prior, in 1977. For those who may not be familiar with the MagCard, it was quite state of the art for its time and was the upgrade to its predecessor, the Magnetic Tape typewriter. Basically, after typing on bond paper, it recorded one page at a time, provided you inserted a magnetic card (much like an IBM card, but Mylar and magnetic) and pressed Record, before turning the machine off—or else it was not saved. Once recorded, the file could be edited by picking the related card to the desired page from the labeled envelope, inserting it into the card reader, and then pressing Read. The file then could be played back one page, one line, one word, or one character at a time. The playback was rather quick, making it easy to miss the mark. Sometimes, the paper ripped during a return motion. This was one of those moments; blame it on user error, but I finally thought, enough was enough, and enrolled at the local community college. This was the beginning of my journey into advanced learning, leading to a BS degree in information management, but definitely not the end of using other machines that would enter NSA’s workspaces. They provided more word processing technologies, office automation, and then advanced further into end user computing.

Maureen McHugh, 1969 I graduated from Marywood College in May, 1969. My experience with computers was limited. I was a Math major in a college whose curriculum focused mainly on training teachers. I did not want to be a teacher. It was obvious in 1969 that computer work would be an exciting field and I could get in on the ground floor. As a senior in college, I took a FORTRAN II class. The teacher was a business professional who taught a few night classes at the college. He had a customized van in which he had a card punch machine, a printer, and sorter. It was only accessible a few hours a week including class time. Writing and debugging our “toy” programs was difficult, to say the least. We submitted our

60

punch card deck to the teacher, and he would return it the following week after running it on a computer back in his office. One turnaround per week! A single typo could set you back two weeks. I think I got a B in the course, but certainly did not feel as though I had mastered FORTRAN.

Peggy Strader, 1969 During my intern tour, I was introduced to the UNIVAC 494 and the SPRYE assembly language. This was an octal-based system so I learned and became proficient in reading dumps in octal. On this system I honed my skills in SPRYE, FORTRAN, and ALGOL. I believe this group was responsible for the first Information Storage and Retrieval System named TIPS (Technical Information Processing System) and its retrieval language named TIPS Interrogation Language. On this system, we were able to put our queries on model 35 teletypes and it would search magnetic tapes of data or magnetic drums for the information requested. I became the user/customer interface for these systems, often teaching the Boolean logic and constructs necessary to retrieve the information needed. Lois Gutman, 1970 I had no real computer or programming experience other than creating small card decks for overnight runs on a cardpunch machine in a summer job at Johns Hopkins University. NSA’s high-level programming language at the time was IMP,13 running on the operating system FOLKLORE, NSA’s homegrown time-sharing system, developed by the Institute for Defense Analyses in Princeton, NJ, USA. Everyone in my office used CDC-6600 computers and sat in a large open tube room, a room full of Cathode-ray tube (CRT) terminals connected to a mainframe where programmers could work on their code in the NSA Headquarters building basement. Operators hung large magnetic tape reels for users. We stacked the reels on our desks (under sheets of black cloth for security) and made hanging decorations from colored plastic write rings. Toby Merriken, 1970 Fresh out of college, I went to work for NSA in 1966. I started out in the Cryptanalysis Intern Program and became certified as a professional in that field. Shortly after that, in 1970, I joined

IEEE Annals of the History of Computing

a newly created branch dedicated to using computer science for the first time for cryptanalytic applications. With no computer training or experience, I wrote programs in FORTRAN and learned a lot on the job. I wrote each program in longhand and took it to a staff of key punch operators to transfer to keypunch cards. I then took the cards to the remote job entry (RJE) room housing a printer and a card reader, into which I manually fed the cards. The input went to the computer mainframe and was printed out on the printer in the RJE room. The computers were high tech for the time but not interactive. A great deal of time transpired between the writing of a program and the implementation of it. I left this branch in 1974 to return to cryptanalysis and eventually became a linguist.

Marie Rowland, 1970 When I started at NSA, I can honestly say I knew nothing about computers. I had gone to an all-women’s college and majored in math. The only exposure to computers we experienced came one afternoon when a guest speaker explained to us that we would all have to learn a language called FORTRAN. So, when I landed at the Agency with my group of new mathematicians, management decided that with my background I should start at the beginning. I was assigned to the Research organization, where they handed me a small box and explained they were doing research on how small computers might one day be used to schedule jobs for large computers. My task was to teach the small box to tell time. Now, I was naı¨ve enough to believe this and did not realize that it was actually a good exercise for me to learn about how to program computers and really understand them from the inside out. I started reading the manuals and some library books. Each morning I plugged in my little box and got it going with a paper tape from a teletype, starting a heartbeat interrupt, a periodic signal that the hardware generated to indicate its working or to synchronize other parts of the system. I could count these beats and get up to a second, then a minute and so forth, and thus tell time. The little machine only had a few instructions such as load, compare, and store so it seemed much easier than the FORTRAN description. The biggest headache was working

April-June 2020

with the teletype and paper tape. An “all thumbs” affliction was to plague me through card decks and keyboards, through all my years of writing code. The library books said I could name my variables anything I wanted. I took this to heart and called them names from the book I was reading, The Hobbit. Thus, “BILBO” became the second counter. Eventually, the person guiding me looked at my work and gently mentioned that it was traditional to name the variables after the function they performed so other people could follow the program. DUH! Later, as I was finishing the project, I asked if he thought I should put in a routine to handle Y2K—something I had discovered in my library research. I do not know how he kept a straight face when he replied that it probably was not necessary for the purpose of this project. I often remembered this Y2K-innocence when it struck with a vengeance years later. As it turned out, teaching this little machine to tell time was a very good introduction to the world of computers. Programming it illustrated the “edge” of the hardware and software divide and left me completely fearless to wade into all kinds of hardware–software issues. I realized what a leap it was from the early computer greats made in the 1940s and 1950s when their research allowed them to move from purposebuilt machines to building machines that kept both the instructions and the data that the instructions worked on in the same form. I went on to write many programs and eventually received an MS in Computer Science from Johns Hopkins, but I was always able to view the complexity of tasks through the lens of my first project.

REFLECTION We all had memorable encounters with our first computers and went on to have rewarding careers in technology. Our Queens of Code are good examples of how women were working with early computer technology. NSA gave us opportunities to excel in this exciting new career field. Over the past 50 years, women have continued to bring their talents and skills to the technology revolution. We hope our project will encourage other women computer pioneers in both the public and private sectors to step forward and tell

61

Anecdotes

their stories. While much has been written on the low percentage of women graduating with computer science degrees14 and problems with retaining female technical employees, it is critical for the future of tech that women’s ideas and points of view be part of future developments. We hope our stories will inspire more women to pursue STEM careers. Look for more stories from the Queens of Code and follow our journey on Facebook: https://www.facebook.com/QueensofCode/ Our website: https://Queensofcode.com On Twitter: @QueensofCode

6. R. Predmore-Lynch. Accessed: 2019. [Online]. Available: https://www.nsa.gov/About-Us/Current-Leadership/ Article-View/Article/1620951/renetta-predmore-lynch/ 7. I was a co-creator of the popular Micro Adventure series (Scholastic 1984, with Ruth Glick). [Online]. Available: http://www.microadventure.net/Home/ About/. We had a talented team of mostly women writers and programmers and some teens to create the series. 1984–1986 8. Dottie T. Blum, Hall of Honor Inductee. 2004. [Online]. Available: https://www.nsa.gov/About-Us/CurrentLeadership/Article-View/Article/1622398/, Accessed: 2019. 9. Read more about flowcharting templates in the Peggy

& REFERENCES 1. Liza Mundy, Code Girls: The Untold Story of the American Women Code Breakers of World War II. New York, NY, USA: Hachette, 2017.

Aldrich Kidwell’s article in the IEEE Annals of the History of Computing (vol. 41, no. 1). Accessed: 2019. [Online]. Available: https://ieeexplore.ieee.org/ document/8667955 10. Involutory matrices can be used in visual cryptography

2. J. Abbate, “Women and gender in the history of

to transform the data. A more detailed example of the

computing,” IEEE Annu. History Comput., vol. 25, no. 4,

Hill cipher algorithm can be found at Accessed: 2020.

pp. 4–8, Oct.–Dec. 2003.

[Online]. Available: https://pdfs.semanticscholar.org/

3. Friedman is the subject of two recent books: G. S. Smith’s, A Life in Code: Pioneer Cryptanalyst Elizebeth Smith Friedman. Jefferson, NC, USA: McFarland & Company, Inc., Publishers, 2017 and J. Fagone’s, The Woman who Smashed Codes. New York, NY, USA: Dey Street, an imprint of William Morrow, 2017. Agnes Driscoll’s mostly unknown life is

9626/7d1194c8beffc8abcf8b142f9870051bdb7c.pdf 11. S. Snyder, Earliest Application of the Computer at NSA (Snyder), 1973. 12. Details on the history of Punched paper tape. Accessed: 2020. [Online]. Available: https://en. wikipedia.org/wiki/Punched_tape 13. Read more about IMP at: Accessed: 2020. [Online].

the subject of K. W. Johnson’s, The Neglected Giant:

Available: http://www.liquisearch.com/

Agnes Meyer Driscoll. Ft. George G. Meade, Center

imp_programming_language https://www.seas.harvard.

for Cryptologic History, 2015. Another little-known early female cryptologist, Genevieve Young Hitt, is the

edu/courses/cs152/2016sp/lectures/lec05-imp.pdf 14. Only 18% of computer science degrees were earned

subject of B. R. Smoot’s, “An accidental cryptologist,”

by women in 2016. Accessed: 2020. [Online].

Cryptologia, vol. 35, no. 2, pp. 164–175, 2011.

Available https://www.computerscience.org/

4. Hidden Figures: The American Dream and the Untold

resources/women-in-computer-science/

Story of the Black Women Mathematicians Who Helped Win the Space Race. New York, NY, USA: William Morrow, 2016. 5. NSA’s Key Role in Major Developments in Computer Science, Part Two, Accessed: 2019. [Online]. Available: https://www.nsa.gov/Portals/70/documents/ news-features/declassified-documents/nsa-earlycomputer-history/6586785-nsa-key-role-in-majordevelopments-in-computer-science.pdf

62

Eileen Buckholtz is a Maryland-based computer scientist and author of 40 books. She received a M.S. degree in computer science from the University of Maryland, a B.S. in computer and information science from Ohio State University, and a Chief Information Officer certificate from the National Defense University. For over 30 years, she had an exciting and distinguished career working on cutting-edge technology for the National Security Agency.

IEEE Annals of the History of Computing

Department: Anecdotes Editor: David Walden, [email protected]

Learning From Prototypes Zbigniew Stachniak York University

& COMPUTER

HARDWARE DEVELOPMENT often involves a succession of hardware prototypes. These prototypes are often discarded once their functionality is tested, performance measured, and their faults detected and analyzed. Occasionally, functional prototypes are used for a short while for demonstration purposes during products’ preannouncements or unveiling to attract the attention of investors and technology commentators. And this is where the life cycle of prototyping typically ends. Fortunately, some computer prototypes survive and end up in museums where they are preserved for research as they may still hide the seeds of the success or failure of both the final products and the firms that embarked on constructing them, of technological breakthroughs and paradigm shifts that were yet to come. York University Computer Museum in Toronto has several prototypes of the MCM/70 microcomputer, which was possibly the earliest computer mass manufactured for personal use. The MCM/ 70 was designed by a Toronto-based electronics company Micro Computer Machines (MCM) in the early 1970s. I have written about the MCM/70 before.1 Yet, some key questions concerning the

Digital Object Identifier 10.1109/MAHC.2020.2987408 Date of current version 29 May 2020.

April-June 2020

computer’s design and introduction to the market remained unanswered until additional prototypes of the computer were acquired by the museum and analyzed. The computer was publicly demonstrated for the first time during the APL V conference held in Toronto on May 15–18, 1973. Before the arrival of the prototypes at the museum, little was known about this historic presentation. Occasional remarks about the demo buried in oral histories gathered by the museum describe with confidence neither the demonstrated hardware, the scope of the demonstration, nor the reaction of the audience to the breakthrough concept of the portable computer for personal use. Another question that could not be fully answered before the resurfacing of the prototypes was how uncertainty in the company’s decision making impacted its shaping and marketing of personal computing. In this article, I describe how the analysis of the MCM/70 prototypes allowed to answer these questions more fully.

EARLY MCM/70 PROTOTYPES MCM built several prototypes of the MCM/70 before the computer’s manufacturing began in mid-1974. All of them have their roots in the KeyCassette concept developed by the company’s co-founder and first president Mers Kutt. A

Published by the IEEE Computer Society

1058-6180 ß 2020 IEEE

63

Anecdotes

Figure 1. Drawing of the Key-Cassette by M. Kutt (1972). Source: York University Computer Museum.

drawing of it can be found in design notes that Kutt kept from late 1971 until mid-1974 (see Figure 1). The one-page sketch depicted a portable computing device with built-in keyboard, one-line display, cassette storage, and acoustic coupler with built-in modem for communication over phone lines. In addition, the KeyCassette was to be programmed in the APL language. The Key-Cassette concept showed several key aspects of personal computing philosophy that MCM would be developing in the coming years—an individual-focused complete computing environment that was easy to learn and interact with. The first attempt at implementing the core Key-Cassette features was a single-board computer put together by MCM’s chief hardware engineer Jose Laraya in mid-1972. His computer utilized an Intel SIM8-01 simulation board which the semiconductor company offered to electronics engineers for experimentation with its novel microprocessor and Eprom devices. Although the Sim8-01 architecture was inadequate to achieve MCM’s design objectives, this first prototype confirmed that building a versatile microprocessor-based computer was feasible.2 Not much is known about the next, rackmounted engineering prototype constructed by Laraya and his team soon after the SIM8-01-based prototype was declared a dead end. It was sufficiently advanced to be demonstrated to shareholders as a proof of concept on November 11, 1972. At the end of 1972, the main design issue faced by the company was an insufficiency of memory: the Intel 8008 microprocessor at the

64

heart of the computer could directly address only 16KB of memory while the APL interpreter alone called for much more than that. Furthermore, the MCM engineers had to find a way to “compact” prototype’s rack-mounted hardware to fit it into a small enclosure planned for the portable personal computer.

PC DEMONSTRATED From the early stages of the MCM/70 design process, MCM used the computer’s prototypes as demonstrators. In April 1972, in his corporate notes, Kutt expressed with some urgency the need to develop a demonstrator by early June 1972. Under the heading “Shortcut to Demo,” he considered packing a power supply and a printed circuit board (PCB) with some MCM/70 circuitry on it into a standard desktop calculator case to have something to show to the potential investors. In the end, MCM came up only with a cardboard mockup which, as it turned out, sufficed to secure venture capital from a law firm in downtown Toronto. MCM built the first functional MCM/70 demonstrator in early 1973, in time for the computer’s public presentation in May during the APL V conference in Toronto. The unveiling of the MCM computer at the conference was a landmark event in the history of personal computing because it showed for the first time that a practical, portable, general-purpose computer designed for personal use and programmed in a high-level language could be economically manufactured.3 Unfortunately, not much is known about that

IEEE Annals of the History of Computing

Figure 2. Wide-case prototype of the MCM/70. MCM promotional photograph (1973). Source: York University Computer Museum.

demonstration. Only a brief statement about the showing of “the first stand-alone APL microcomputer which elicited a great deal of interest” can be found in L.B. Moore’s conference report published in APL Quote-Quad.4 Furthermore, none of the former MCM employees interviewed by me could describe the demonstrated hardware or software with confidence. That changed in 2017 when York University Computer Museum obtained one of the MCM/70’s prototypes and a portfolio of early MCM/70 design documents. When analyzed, these objects helped not only to identify the demonstrated prototype but also to determine, in general terms, the presentation’s content. Among the donated documents, there are two drawings of PCB layouts. The first of these drawings is dated May 9, 1973, and titled “MCM 70 PROTOTYPE.” The second drawing, dated June 26, 1973, has reference to neither a prototype name nor a revision version. Both drawings define single-board computers, i.e., computers whose main circuitry reside on a single board. It is highly likely that the computer demonstrated by MCM at the APL V conference was the so-called “wide-case prototype” that the company extensively used for promotional purposes in 1973 (depicted in Figure 2), and that the computer’s hardware was defined by the June 26th documentation. Here is

April-June 2020

why—some early documents sent by MCM to its shareholders stated that several MCM/70s were constructed and ready for field trials in early May (hence, before the APL V conference), and that by August, MCM had built ten wide-case machines “for marketing test purposes.” One of these documents also includes an image which depicts exactly the same computer as the wide-case prototype shown in Figure 2.5 Because of MCM’s limited manufacturing capabilities and the fact that the design and manufacturing of the wide-case prototype’s case took considerable time, the production of the ten prototypes would have had to begin before May. It is therefore reasonable to conclude that the MCM computers available in May were the first of the ten destined for field trials in August. Hence, what MCM demonstrated during the APL V conference was one of the wide-case prototypes available in early May. To establish the hardware makeup of the demonstrated MCM computer, I compared the June 26th PCB drawing with the published specifications of the wide-case prototype and with its image (see Figure 2). From the published dimensions of the wide-case prototype and those found on the drawing, it is evident that the PCB depicted in the drawing would fit perfectly into the prototype’s case. Furthermore, the types,

65

Anecdotes

locations, and dimensions of the keyboard, numeric keypad, plasma display, and switches (as shown in Figure 2) match the PCB’s layout exactly. Therefore, the wide-case prototype was most likely the single-board computer defined by the June 26th documentation. It was built around the Intel 8008 CPU and was equipped with 2KB of RAM and 10KB of ROM which contained, among other software, an interpreter of a dialect of APL—the MCM/APL. Having the hardware identity of the demonstrated computer established, we can turn our attention to the question of the demonstration’s content. In a 2003 interview, Gord Ramer, who implemented the MCM/APL for the MCM/70 computer recollected that “The demo [computer] had almost the complete APL implementation. There were still bugs to be avoided [. . .] We did not have a special demo version of the software, so the demo was tightly controlled, i.e., the person on the keyboard new what to avoid.”6 APL programming language offers a range of built-in functions and MCM most certainly demonstrated the capabilities of its computer by executing programs involving some of them. But which ones? Early MCM/70 promotional documentation frequently included computer code comparisons to demonstrate the efficiency of coding in APL as opposed to programming in languages such as Fortran and Basic. Computing the average of a set of numbers (APL code (þ/X)rX ⎕) and sorting a set of numbers (APL code X[⍋X ⎕]) where given as examples. But these programs would be considered rather elementary by an APL programmer who, instead, would rather see examples involving APL functions whose evaluation required considerably more time—functions such as matrix inverse which was frequently used for benchmarking. Unfortunately, the low speed of the Intel 8008 processor inside the MCM/70 demonstrator meant that, most likely, MCM demonstrated programs with short execution times and only those that could be directly entered via keyboard. Indeed, the demonstration of long multiple-line APL programs would require a transfer of such programs from external storage into the computer’s memory. In a preannouncement document released in May 1973, MCM listed an external cassette unit as an available option. In principle, such a unit could have been

66

interfaced with the prototype via the computer’s communication bus called Omniport. However, it is doubtful that such cassette storage was available for the wide-case prototype at the time of its presentation in Toronto. The computer could only be equipped with up to 2 KB of RAM and that was not enough to support an external cassette unit and to provide enough memory for the storage and execution of user’s applications.7 In the end, the announced cassette unit did not materialize as a commercial product. Instead, MCM offered three production models of its MCM/70 computer: a model without cassette storage and with 2 KB of RAM (as in the wide-case prototype) and two models with built-in cassette drives—a single cassette model with 4 KB of RAM and a twocassette model with 8 KB of RAM (see Figure 4). Scarce primary sources detailing the presentation make it difficult to ascertain the APL community’s response to the demo and to MCM’s concept of a personal APL computer. According to Moore’s above-quoted statement, the MCM/ 70’s presentation was met with interest. In a 2001 interview, Mers Kutt made a similar comment noting that the computer astounded a group of IBM employees attending the conference. In fact, two of the first dozen MCM/70 units manufactured by MCM went to IBM’s General Systems Division in Atlanta, Georgia, for the purpose of “research and analysis.” In 1975, IBM announced its own APL desktop computer—the IBM 5100.8 However, other MCM employees assisting with the presentation recollected a less enthusiastic response from the audience who did not give the MCM/70 computer too much serious thought. Joey Tuttle, a former IBM employee and a member of the IBM 5100 development team, recalled a similar presentation of the IBM 5100 at IBM Rochester. During the presentation, “one of the attendees shouted out, “run þ/40000 r 1”” which would create an array of length 40 000 filled with 1s and to return the sum of all the elements stored in the array—a simple way to test memory. When the presenter entered the expression and the computer seemed frozen for an extended period of time, “the anticipation and the increasingly nervous body language of the presenter, turned to laughter.”9 In the end, the demonstrated IBM 5100 took over 3 min to finish the evaluation of the expression in question, which

IEEE Annals of the History of Computing

was very slow in comparison with the state-ofthe-art IBM mainframes running APL. It is plausible that similar requests were made during the MCM/70’s demonstration and that the slow speed of the computer elicited similar reactions to those recalled by Tuttle. Indeed, a request to evaluate an expression such as 0.7i255 to generate consecutive 255 integers and to divide each of the numbers by 0.7 could easily have resulted in a snicker as it would have taken the machine approximately 50s to evaluate it while the execution of the same code on an IBM System 360 mainframe would have produced the output in a fraction of that time. “Such audiences can judge harshly,” concluded Tuttle.

EXECUTIVE—THE MISSING LINK In the summer of 1973, several MCM/70 demonstrators were on their European and North American promotional tours. Most were the wide-case prototypes. But one was an altogether different piece of hardware, assembled in record short time for a demonstration at the APL Congress that was to take place in August at the Technical University of Denmark in Lyngby, north of Copenhagen, Denmark. In July 1973, MCM decided to pack the current MCM/70 hardware, including a keyboard, a one-line plasma display (Burroughs SelfScan), and a single cas case and to operate sette drive, into an attache all of it on batteries exclusively. The company expected considerable marketing gains from the planned ground-breaking presentation of a never-seen-before luggable, battery operated, general-purpose computer which they named the Executive. The gamble played off. On August 23, the day after the Executive’s demonstration, the Danish daily Politiken published a front-page article about a sensational computer from Canada. The article included two photographs depicting MCM employee Ted Edwards operating the Executive on the doorstep of the auditorium where the conference took place (see Figure 3). “In all modesty, a real sensation occurred yesterday when the International APL Congress was about to begin at the Technical University of Denmark.” wrote Politiken. “When the buses with conference participants from 24 countries arrived from hotels in

April-June 2020

Figure 3. This photograph of Ted Edwards demonstrating the Executive during the APL Congress in Copenhagen appeared in Politiken on August 23, 1973. Source: York University Computer Museum.

Figure 4. Two-cassette MCM/70 (model 708). Source: York University Computer Museum.

67

Anecdotes

Copenhagen, he [Edwards] was sitting on the steps of an auditorium building solving complex problems on his data processing machine placed on his lap. Several experts who went by thought that the computer was a joke. Many stated that such a machine was impossible. But others who followed Ted Edwards’ programming computations [. . .] admitted that the machine performed exactly as considerably larger IBM computers running APL and connected to an electrical outlet. There just was not any power cord attached to Ted Edwards’ briefcase.”10 A look inside the Executive’s briefcase enclosure gives the impression that the computer’s hardware was packed in a rush using, what would be best described as a “chain saw and duct tape” approach.11 The computer’s keyboard had been crudely sawn-off from the widecase prototype’s main board. Other components were attached using means ranging from nuts and bolts to sticky tape. The large ROM board consisting of 88 Intel 1702A Eproms that barely fit inside the case was placed at the bottom and, it seems, was kept in place only by means of the numerous wires pressing against it. The built-in cassette drive allowed for loading and storing APL programs and data. However crudely built, the analysis of the Executive revealed much about the evolution of the MCM/70 concept. The single-board approach employed in the widecase prototype was replaced by a modular architecture in which computer hardware was logically divided into interconnected individual modules such as CPU, ROM, RAM, and interface modules that occupied separate circuit boards and communicated over a common bus. The Executive’s CPU and cassette interface boards were early versions of the boards that would eventually populate the production model of the computer. On the other hand, the RAM board (4 KB) was, possibly, the one developed earlier for the rack mounted prototype and would be redesigned for the production model. Finally, the large ROM board inside the Executive was the same as the one used in the early production models. Because in early 1974, ROMs containing MCM/70’s systems software were not yet delivered by the supplier (Electronic Arrays), the first MCM computers were equipped with the same ROM board found in the Executive and, because

68

of its large size, it was mounted externally under the bottom of the computer’s case (thus earning it a nick-name “pregnant bottom”). This shows that from the hardware architecture point of view, the Executive was almost identical to the production model and, hence, that by July 1973, the company had already abandoned the singleboard approach represented by the wide-case prototype. For a brief time, MCM contemplated manufacturing the Executive and made announcements to that effect in the press. In October 1973, the computer was shown during the National Computer Show and Conference in Toronto but, in the end, the company quietly abandoned the Executive concept. The surviving MCM corporate documents do not provide any justification for the decision. However, some arguments against the Executive can be reconstructed by analyzing its design and another prototype put together by Edwards some years later. To begin with, the Executive was an attempt at building a fully luggable, battery-operated, personal data processing system. It was to offer complete data processing and communication functionality at the user’s fingertips. However, most if not all of the Executive’s features were planned for the MCM/70 desktop which, in addition, was not much larger or heavier than the Executive. Like the Executive, it could be operated on batteries and carried around in an elegant leather case that could be purchased for $150 from MCM. Most likely, there were also arguments of a technical nature against the Executive. Some of these arguments can be deduced from the Executive-like computer designed and built by Edwards around 1975–1976 (see Figure 5; I will refer to this computer as the Executive-E.).12 From a hardware point of view, the Executive-E is an MCM/70 clone. All the PCB boards used in it are the same as those found in the production model of the MCM/70 or its MCM/700 refinement. However, from a packaging perspective, it is the Executive repackaged into an all-aluminum case. The addition of the second cassette drive may suggest that one of the issues with the Executive was its limited memory and, as a consequence, limited functionality. While the Executive used its single cassette drive for external storage, the amount of RAM that could be made available to a user remained low.

IEEE Annals of the History of Computing

1001–in 1981. In the following years, several GRiD laptops would find their way into space supporting NASA’s early Shuttle missions.13

MARKETING PERSONAL COMPUTING

Figure 5. Ted Edwards’ Executive-E. Source: York University Computer Museum.

MCM solved the insufficient RAM problem by developing a cassette-based virtual memory system (called AVS) that made over 100 KB of memory available to applications. But AVS required a dedicated cassette drive and there was simply no space left in the Executive’s case for the second drive. There was no space left for anything at all, not even an internal fan, which brings us to the second issue—heat dissipation. The Executive’s case had no provisions to dissipate heat during the computer’s operations. In the MCM/70 production model, which also lacked an internal fan, the heat dissipation problem was solved by packaging all the boards in special aluminum casings sandwiched together to form the back end of the computer and to act as a giant heat sink. In the Executive-E, Edwards used the entire aluminum case of the computer as a hit sink. It is likely that the Executive-E was an attempt by Edwards at salvaging the Executive concept. Because of its sturdy packaging it could have been marketed as a luggable data collection and processing system for use in harsh rugged environments. However, by the time its design was finished, Edwards was no longer with MCM. It would be left to another company to build and successfully introduce a portable battery-operated computer for rugged environments. GRiD Systems, Inc., introduced its first laptop–the GRiD Compas

April-June 2020

With the MCM/70 personal computer concept, MCM hoped to create a lucrative niche on the electronics market between the inexpensive, easy-to-use desktop electronic calculators that were inadequate for many data processing tasks, and large, expensive, and complex to operate general-purpose computers. To succeed, the company needed to support its personal computing venture with a well-defined and appealing product philosophy and attractive marketing to promote it. MCM promotional documents from the mid– 1973 paint an interesting picture of how the company was stitching together a personal computing paradigm from the computer’s basic features as well as its hardware and software options. An August 24, 1973 portfolio of MCM/ 70 preannouncement documents distributed to MCM shareholders already contained an outline of the paradigm. The revolutionary concept of the MCM/70 is that it brings to the world of computing what the $100 hand held calculator brought to the world of calculating. [It is] of a size, price, and ease-of-use as to bring personal computer ownership to business, education, and scientific users previously un-served by the computer industry.14

But that’s not all. The portfolio also listed the MCM/70’s basic features together with available options. The list was a very long one: two types of printers (one external and one built into the computer’s chassis), cassette drives, serial communications interface, CRT (Cathode Ray Tube) display, RAM expansion of up to 64 KB, and even plug-in ROM modules with preprogrammed applications software. These features and options were designed to capture the core features of the personal computing environment that MCM wanted to offer with its desktop. However, throughout 1973 and 1974, the list of the MCM/ 70’s available options was a moving target, continuously revised with some options added, others dropped to reappear again. Were these changes reflective of uncertainty in company’s decision-

69

Anecdotes

making or of an evolving point of view on personal computing that necessitated the shifting of priorities for the development of at least some of these options? Or were they part of a marketing ploy to overwhelm potential customers and MCM shareholders with claims of its offering’s sophistication and completeness, an early example of vaporware? The analysis of the surviving MCM hardware and corporate documents shows that all of these concerns were the case. In late 1973, MCM dropped the preprogrammed ROMs option all together, gave up on memory extension to 64KB, replaced the built-in printer option (which had never materialized) with an external impact printer. But not all of the company’s claims were a marketing gambit. The Executive’s hardware already featured a partially implemented communications bus (Omniport) that allowed the computer, in principle, to interface with printers.15 However, what the company did not have at that time was the printer itself, and it was not until mid-1975 when, finally, it offered its MCP-132 printer (the Diablo HyType I printer). At that time, MCM also upgraded the MCM/70 with an EAI communications interface that allowed the computer to communicate with a range of other peripherals. MCM’s stated objective “to offer a complete micro computer system” was finally realized.16 The MCM/70’s built-in display (Burroughs SelfScan) was one of the core features of the allin-one hardware concept. While such a display enhanced portability, the single-line 32-character-long solution adopted for the computer offered a convenient display environment for only rudimentary APL calculations. MCM maintained that MCM/70’s display was sufficient for many tasks because of APL’s coding efficiency. For all other applications, there was to be a CRT display, already listed in August 1973 as an available option. Of course, the company could have equipped the MCM/70 with a multiple-line display, as was done for some calculators of the era, if it were not for the fact that the computer required a dedicated segment of RAM for storing characters to be displayed (display memory). More display lines would take away precious RAM from the user’s work space. The RAM shortage was so severe that display memory was also used for temporary storage during computations.

70

A first-time user would have been quite surprised and astounded when after entering a command, such as the above mentioned 0.7  i255, the computer would begin a 50-second-long display of strange and dazzling patterns (i.e., visual representations of temporal data stored in display memory) before eventually clearing everything and showing the result of the computation. Therefore, having a single line display built-in and an external CRT terminal as an option seemed like a reasonable solution. But the CRT option disappeared from MCM’s 1974 promotional literature and was not discussed (as a work in progress) in any surviving minutes from MCM managers’ meetings. It would not be until 1976 that MCM finally introduced such a display for its new computer—the MCM/800. This might suggest that the announcement of a CRT display as an “available” option made three years earlier was, to put it mildly, a careless instance of the company “snowing” customers and shareholders with options that were only meant to enhance the MCM/70’s image. Or is it? To answer this question, I analyzed the ROM boards inside the MCM/70, Executive-E, and MCM/800 computers. The inspection of these boards showed that the same three ROM sockets (out of 16) were left unpopulated in the MCM/70’s and the ExecutiveE while similar boards inside the MCM/800s had all 16 ROM chips installed. It turns out that these three empty sockets were reserved for correcting and future expansion of the MCM/70’s systems software, including the addition of CRT support. One may therefore conclude that financial or other corporate difficulties that MCM experienced in 1974 forced the company to drop several options, including an inexpensive printer and a CRT display, in order to concentrate on the MCM/70’s prompt introduction to the market. However, as evident from the MCM/70’s ROM board, some necessary hardware provisions for such options were made at the start.

LOOK OF A GADGET The MCM/70 concept was shaped by ideas coming from several sources, one of which was the calculator industry. From the late 1960s, desktop programmable calculators were presented as minicomputers and even as minis for personal

IEEE Annals of the History of Computing

use, as was the case with the Olivetti Programma 101 introduced in 1965. Furthermore, the manufacturers of hand-held calculators marketed their devices as consumer electronics gizmos. The MCM/70 was to be both; a stylish “gizmo” look was regarded just as important a feature of the MCM/70 as its conception for personal use. Through the industrial design of the MCM/70 the company wanted to draw attention to a new computing paradigm and to create its own unique identity. In his 1965 article “The Great Gizmo,” Reyner Banham characterized a gizmo as

2. For more information on this prototype, see Z. Stachniak, “SIM8-01: A proto PC,” IEEE Ann. History Comput., vol. 29, no. 1, pp. 34–48, 2007. alisations 3. Three months earlier, a French company Re   Etudes Electroniques (R2E) announced its microprocessor-based Micral (see, Bui, R. Un miniordinateur pour moins de 8500FF, zero.un. informatique hebdo, no. 228, Feb. 12, 1973, pp. 1 and 5.). However, the Micral was not intended to be a personal computer: “MICRAL’s principal use is in process control. It does not aim to be an universal minicomputer,” Micral User’s Manual, R2E, Jan. 1974, p. 66. 4. L.B. Moore, A report of APL V Conference, APL Quote-Quad, Jun. 1973, pp. 20–21.

[. . .] a small self-contained unit of high performance in relation to its size and cost, whose function is to transform some undifferentiated set of circumstances to a condition nearer human desires. The minimum of skills is required in its installation and use, and it is independent of any physical or social infrastructure beyond that by which it may be ordered from a catalogue and delivered to its prospective user.17

5. The documents in question are dated May 4 and Aug. 24, 1973. York University Computer Museum, MCM Collection. 6. Author interview with Gord Ramer, Jan. 15, 2003. 7. A large portion of the computer’s 2 KB of RAM was used for display, APL execution stack, and a variety of tables required by the computer’s operating system. Interfacing a cassette drive with a computer would require an allocation of additional RAM to store

As early as the MCM/70’s prototyping stage, the company’s description of the computer provided an almost undeviating instance of Banham’s characterization of a gizmo. The Key-Cassette concept was a gizmo (although only on paper), and so was the Executive. The design of the wide-case prototype accomplished not only the requirement of hosting all the necessary hardware in a single box but also of finding the right balance between stylish eye-catching design, functionality, and practicality. Two years of design experiments culminated in the 1974-release of the MCM/70’s production model which inherited the all-in-one concept from the Key-Cassette and the Executive, and a defining stylish design from the wide-case prototype. In the late 1970s, the nascent home computer industry would follow in the footsteps of MCM choosing appealing, dashing designs over the industrial look of minicomputers.

information about the interfaced device and the tape’s content leaving little space for anything else. 8. I have encountered no evidence for the design of the IBM 5100 being, in any way, influenced by the MCM/70. 9. J. Tuttle, private communication, Jan. 2020. 10. Politiken, Aug. 23, 1973, pp. 1 and 20. 11. The Executive was donated to York University Computer Museum in 2017. 12. The date codes found on its ICs, suggest that the computer was put together around 1976. Because Edwards left MCM at the end of 1975, he continued the design after leaving MCM. 13. Accessed on: Mar. 2020. [Online]. Available: https:// airandspace.si.edu/node/35305 14. MCM/70 pre-announcement shareholder documents, Aug. 24, 1973. York University Computer Museum, MCM collection. 15 This was confirmed using the MCM/70 emulator developed at York University Computer Museum and which uses almost identical systems software to that found in the Executive’s ROMs.

& REFERENCES 1. See Z. Stachniak, Inventing the PC: The MCM/70 Story, McGill-Queen’s University Press, 2011, and Z. Stachniak, MCM on Personal Software, IEEE Ann. History Comput., vol. 39, no. 1, pp. 29–51, 2017.

April-June 2020

16. MCM Newsletter, no. 1, 1976, pp. 2 and 3. The MCP-132 printer plotter was sold by MCM for $4500 which was almost the same price as the MCM/70 in its basic configuration ($4970). 17. R. Banham, The Great Gizmo, Ind. Des., vol. 12, pp. 48–59, 1965.

71

Department: Interviews Editor: Dag Spicer, [email protected]

Interview With Michael R. Williams Abstract—In this interview, Mike Williams discusses his college career in Calgary, Alberta, Canada, where he first encountered computers, and in Glasgow, Scotland. In graduate school in Glasgow, he first became interested in computing history. He spent the rest of his professional career back in Calgary in the Department of Computer Science at the University of Calgary from which he retired many years later. In parallel with his academic career, Mike was involved in a variety of computing history activities with other institutions and with other computing historians. He also served in many positions with the IEEE Computer Society and the IEEE, including as 2007 President of the Computer Society.

& NOTE:

THIS IS based on an interview conducted under the auspices of the IEEE. The full text is available here: https://history.computer. org/leaders/mike-williams.pdf Walden: Today is February 23, 2014; I’m Dave Walden and I’m with the 2007 Computer Society President, Michael R. Williams. Please tell a bit about where you’re from, your youth, your hobbies; anything you think is interesting about your parents and growing up. Williams: I was born in Calgary, Alberta, Canada, in 1942 to a family that was very much a working-class family. There was never enough money to go around. I had two older brothers and a younger sister, and it was a pretty happy arrangement. I went to school in Calgary and by and large, stayed in Calgary for most of my life. There were plenty of times I’ve been elsewhere,

Digital Object Identifier 10.1109/MAHC.2020.2992827 Date of current version 29 May 2020.

72

1058-6180 ß 2020 IEEE

sometimes for several years at a time, but I always tended to come back. I was always interested in science, and I read quite a bit about science, and of course, science fiction as well, and quite enjoyed a chemistry teacher in high school, who was very good and encouraged us to do all sorts of things that weren’t in the curriculum and even supervised when we were doing some fairly dangerous things, like creating bromine gas in a fume hood. There were half a dozen of us, I guess, in this chemistry club and we really got into doing some interesting stuff. So it was only natural that when I wanted to go to university I would’ve taken a degree in chemistry. After about two years I began to realize that chemistry was not so exciting as I once thought it would be and I had been introduced to computers at that point, as the fledgling university had acquired an IBM 1620 [data processing system], and I had known somebody who was a school chum, who introduced me to computing. However, I would’ve had to sort of go

Published by the IEEE Computer Society

IEEE Annals of the History of Computing

back and catch up all sorts of courses I didn’t have, so I actually ended up having a degree in chemistry, which I never used. I certainly had long since decided that I didn’t want to pour gunk from one test tube into another for the rest of my life and I was actually employed at the university, doing some programming and things for both the administration and some of the research staff. So I got very interested in computers and how they worked. Walden: On the IBM 1620—you were doing assembly language or FORTRAN [programming]? Williams: Both, but mainly assembly language because of the things I was doing for the university administration, scheduling final examination timetables, sectioning students to classes. When I finally graduated, it was a bit of a tough time because there was no easily available degree in computer science anywhere. So I took what money I could make over the next summer, and put it in my jeans pocket; did the usual student thing: I got a boat across the Atlantic Ocean and ended up in Egypt. And Egypt has been one of my favorite places on the face of this earth ever since. It’s a spectacular place. Even in later years, I took courses from the archaeology department on reading Egyptian hieroglyphics and things like that. I was wandering around Alexandria when I discovered a shop where I could buy a magazine called The New Scientist, an English language magazine. And being short of English language reading material, I immediately bought it, sat down on a rock on the beach, and started looking through it. I eventually discovered an ad in the back for the University of Glasgow in Scotland that said they were looking for somebody to do some programming on their computer. I thought to myself, well, that would be interesting but this is a long way from Scotland and I didn’t do anything about it. But I eventually started to run short on money and made my way hitchhiking, and things like that, back through Europe and into Britain. I had earlier met a fellow from Edinburgh, a student, and I thought oh, I’ll hitchhike my way up to Edinburgh and I can sleep on his couch or something. When I started hitchhiking, my ride actually went to Glasgow, not to Edinburgh. When I got to Glasgow, I put up at the Youth Hostel, and it was raining cats and dogs. So I thought I’m not going to go out hitchhiking to get to Edinburgh, and I asked

April-June 2020

the warden what there was to do, and he said the university’s across the way, you can go and read in the library or something. And that’s when I thought oh, yeah, this was the place that had the ad for the computer. So I went over to the university and wandered around until I found out where the computer was and talked to a few people in the hallways, graduate students and things like that, some undergraduates; and then just left—went to the library. It was closing early for some reason that day, and they rang a bell at five o’clock to throw everybody out. I walked out intending to go back to the Youth Hostel when I noticed a fellow that I had talked with that morning named Sa’ad ben Hamid, from Libya, who was a graduate student. He noticed me at the same time and he said “oh, weren’t you the one who was down and talking to us?” He said “we told our prof about you at lunchtime, he wants to talk to you.” I said “good, I’ll go ‘round, see if he’s there.” “Oh no, no, no; he’s gone home.” I thought okay, well, it’s still raining, I’ll see him tomorrow. When I saw him, he was sort of interested because I had been talking to some of these people about the stuff I had done in Calgary earlier, the work for the Registrar and things. He obviously had picked up on that somewhere, and by sheer coincidence the day before, the University of Glasgow Registrar had come to him and said “do you think this computer might be useful to us here in Glasgow?” Dennis Gilles, who’s the prof in question—the only full professor and Head of the Department, said “let me think about it.” So after talking to me for a couple of hours, and finding out what we were doing with the computer in Calgary, he finally said “do you want a job?” And I said “yes, but only if I can do graduate work in computer science at the same time.” And he said “done!” I said “don’t you need transcripts and things to show I’ve got a degree?” He said “we’ve been talking for two hours, I know you had lots of experience and I believe you’ve got a degree. You’re in.” I ended up staying there for four years and completing a Ph.D. in timetabling for universities, classes, schools, and things like that. Walden: When you started in this area at Calgary, were you reading the literature about scheduling things or were you making it up yourself?

73

Interviews

Williams: We were making it up ourselves because there was very little literature at that time. We just did it ourselves one way or another. And I think that the same program for scheduling final exams was still operational in the 1990s. It had been rewritten half a dozen times for different machines, but it was still the same program. Then I think in the 1990s somewhere, they finally farmed it out to some commercial firm. So it had a good long life. Walden: Back in Glasgow, then, you were doing scheduling of things? Williams: Yes, only at this point, I was getting much more into the theory of graphs and whatnot; and doing an awful lot of reading. So my thesis was graph theory with some practical applications. And now, of course, if I look at my thesis, it looks completely ridiculous. The things that we were trying to do are nowadays probably done in second-year assignments for computer students. But in those days, it was a lot more cutting edge. It was actually more interesting for me because my supervisor, Dennis Gilles, contracted tuberculosis. He was in isolation in the hospital for many months and then recuperating at home for many months afterward. And so it was about another 18 months where he wasn’t there at all, and only minimal time for the next six months. So basically, I was on my own with the help of other graduate students—we tended to help ourselves—and it was, of course, in the 1960s, a very, very small department. Walden: What computer were you using in Glasgow? Williams: It was an English Electric KDF9, a giant machine compared to the IBM 1620 that we had in Calgary. I still rather like that machine— you could write most of your programs in Algol 60—and it was an interesting machine with an unusual architecture. Walden: Was that Brian Randell’s Algol 60 running on that machine? Williams: Yes it was, and Brian and I became friends after a while. Walden: You knew him? Williams: Oh yes, very much so; and we still correspond on occasion. There were several different Algol compilers, actually, but certainly, his Algol was the one that we would debug programs on because it was a very fast interpretive system. The students used that, as well, because,

74

although it was slow in executing, it was very fast in checking syntax. They were creating the curriculum at the time, and the fact that the professor was no longer there proved to be very difficult. One of the senior university people—I can’t even remember who it was, now—held a meeting and said there is a senior lecturer, two other lecturers in the department, and half a dozen graduate students. You people will have to run the department, and you people have to do all the teaching, and you people will have to do everything, even creating budgets and answering correspondence while the professor is no longer capable. And so I got thrown in the deep end. We had to establish the curriculum for the next couple of years, and do the lecturing, and all sorts of things like that ourselves. We organized our own seminars, both amongst ourselves and inviting other people to come and speak. I don’t remember him being there, but it would be likely that we invited people like Brian Randell. That stood me in very good stead because when I finally graduated, I was basically quite independent in research and knew an awful lot about the administration of a university. I had actually tried to find a job back in Canada, and one of the things that I did was to write to my old professor in Calgary, John Peck, and say that I didn’t have any contact with people anymore and did he know if there were jobs going in Canada? I never heard back from him, which wasn’t atypical as John was very bad at answering letters and things. About three weeks later I got a telegram from the Dean of Arts and Science at the University of Calgary, saying that I was appointed, starting September 1, in the Department of Mathematics (which taught computer stuff in those days), and asked me to please confirm my acceptance. Well, this came as a complete surprise and, in fact, I couldn’t come because my Ph.D. orals were going to be at the end of October. So I telegrammed back thanking them for the offer but that I was very sorry but I was committed to staying in Glasgow until about November 1. A few days passed when another telegram arrived saying that arrival November 1 is acceptable, please confirm. I thought, well, I’ve got to have a job to go to, I’ll just take that offer and go to Calgary, and then I’ll work there for a year and find something better. I just left my office in the middle of

IEEE Annals of the History of Computing

December this year and so, in fact, it was 45 years that I was in Calgary, having only intended to stay there for a year. It’s treated me well. Glasgow has [also] been very good to me. They have actually given me an honorary Doctor of Science degree just a few years ago; for work in history. Walden: So at the University of Calgary when you went back for the first year, you weren’t working in a computing technical area, but you were teaching computing courses? Williams: Yes. Walden: And were beginning to work in the computing history area? Williams: Correct. I never really thought I would work in the computing history area. I was certainly interested, but the job at Calgary was technically in the Department of Mathematics. The Department of Mathematics was quite large. They were just discussing dividing the administration into four divisions: pure math, applied math, statistics, and computer science. Each of these groups was to have a chairman, along with the head of the entire Math Department. I was very junior at this point; I was 25 years old, I suppose; and there were some very much more senior individuals, in age. I think by that time, I was the only one who actually had the degree in computer science because they were rare birds, in those days. And so, after the first year, summer came along. My wife had friends in California she wanted to visit. So we took our vacation; went to California; and when we came back a few weeks later, there was a notice waiting for me in my mailbox saying the Dean would like to see me. I thought, “oh dear, I wonder what I’ve done now.” I went and talked to the Dean and he described how the divisions were being set up, because they hadn’t actually been set up yet, and they all needed chairmen. He said he wanted me to be Chairman of the Computer Science Division. I reminded him that I was the most junior person there. And he said “yes, I don’t like the senior ones, I want you to do it. I think you had some sort of administrative experience in Glasgow, didn’t you?” I agreed. He said “good, then you can be Chairman.” I said” but”. . .; and he said, “you don’t have tenure, you have a choice. You can either resign now or you can become Chairman of the Computer Division.” [Laughter]

April-June 2020

Now, I was too green to realize that he probably was joking but I swallowed hard and said, okay. I was only 25 years old and having been there for only five or six months at that point, I ended up chairing this small group of computing people that eventually grew, of course, into its own separate department. But that, too, gave me a lot of interesting administrative experience because we had to hire all sorts of people. Those were the days when money was flowing quite freely so we had lots of money to hire people and buy equipment. People were hard to find, money wasn’t. Now it’s the other way around. Walden: So during this time, you’re teaching courses, administering the Department. . . were you able to do any computing history during this time? Williams: Not very much. But what I did discover was that when I was lecturing, if I just said this is the way it’s done; you know, the parameter passing mechanisms in languages or something; some students would just accept it and go away without ever asking questions or anything else. And other students would not understand and ask “why is it done that way?” If I could step back and say, “look, years ago it was done this way and that didn’t work very well because. . .; so it got changed to this. Then afterward, when a better idea came along, it got changed again and this is what we finally have today, which seems to work fairly well but, you know, possibly more changes will come up.” So giving a more historical perspective to some of these things seemed to help a lot of students understand why we did things, rather than just present it as a God-given situation. And it was things like that, that gradually got me to talk more and more about history and less and less about other things, although I continued to teach programming languages and similar topics for most of my teaching career. I did eventually start a course in the history of computation, and it was the history of computation—not the history of computers—because we went all the way back to ancient Egypt to see how they did their arithmetic, followed up through the years and finally, after about four months of lecturing we’d come to about where the IBM System/360 was developed. In those days that was the point where I’d usually quit because often the students could follow the developments themselves from that point on.

75

Interviews

Although today, they seem to think no computer existed before the Apple. Walden: After you’d been at Calgary for a few years, you were Visiting Professor at St. Andrews in Scotland. Were you just there for a sabbatical year, or six months? Or you just visited occasionally? Williams: I was there for a sabbatical year. That occurred after my term as Chairman of the Computer Science Division stopped and I was given a sabbatical at St Andrews. Primarily it was done this way at the University of Calgary so you were not influencing the next person who came along. Walden: I see, so they wanted you out of town. Williams: Yes, and it’s a good idea because a new person taking on the job should be able to make his or her own way and not be constantly looking over their shoulder or expecting you to show up raving and ranting in their office, or something. So I had a sabbatical at St. Andrews, where I was first able to really get into some history work because they had a very, very good, very old library dating from the late 1400s, I think. Walden: I see from your list of publications that it was just after you began that sabbatical year that you began to have publications in computing history. Williams: There was no place to publish it in those days. Any of the computing journals, like the Communications of the ACM, or the British Computer Society Journal, things like that, they weren’t really into publishing any historically oriented things. In fact, the Communications of the ACM, when I had sent them something, just turned it down flat, saying we’re not interested. I think it appeared in the British Computer Journal eventually. It was shortly after that, that the very first conference on the history of computing was organized at Los Alamos, New Mexico, 1976. It was by invitation only and Nick Metropolis had organized it at his place at the Los Alamos National Laboratories. He had invited computing pioneers to come and talk about what they had done, and it was all recorded on videotape. Really, it was to be a way of just preserving some of this historical information. I plucked up my courage, wrote to Nick, and said I was very interested and could I have an invitation please. He was very polite and

76

generous and said certainly, come along. So I went down there and I got to actually meet most of the early pioneers, one way or another, and that really sparked my interest in a lot of this stuff. It was interesting that you could easily strike up a conversation. For example, I remember sitting beside Konrad Zuse one night and we spent about two hours over a very leisurely supper discussing his work. That opportunity just doesn’t come along very often. Brian Randell was there; he did some historical work on the code-breaking British Colossus machine. He brought somebody named Dr. A.W.M. “Doc” Coombs, who had actually worked on building the [wartime] Colossus machines, and we had sessions where we could ask him things. The British Official Secrets Act was still in force and it’s very draconian, you could go to jail for a very long time by breaking that. Doc Cooms was willing to talk about anything and everything. Brian would occasionally lean over, touch him and say, remember the Act, and Coombs would shut up. We were all ready to kill Brian over it.y Walden: As you moved into history as your research area, was there a particular area you were focused on? Williams: It was basically what was interesting to me at any time. I had started to gather as much information as I could about the early computers, generally vacuum tube one-off devices. There were papers that were published about these machines but they didn’t contain much of the hard detail. For that you had to actually ask somebody who knew. The published papers didn’t tell you much in the way things like how the circuitry worked and it didn’t tell you the secrets of how you wired this thing up to do the work. It was usually just very hand-waving kind of description. Having met these people, it allowed me to ask the more interesting questions—at least to me,—and I developed the technique of sitting and talking to people, and then running back to my room, if I was out at a conference or something, and scribbling like mad for the next hour to try and get all these things down on paper so that I didn’t forget. It was that way that I talked with people who had, for example, worked with Alan Turing. In those early days, nobody was willing to talk to me about y

See: Williams, M.R.,“The First Public Discussion of the Secret Colossus Project,” IEEE Annals of the History of Computing (vol. 40 , Issue: 1 , Jan.–Mar. 2018)

IEEE Annals of the History of Computing

Turing, and why he killed himself, and things. This was just not done. I can remember one fellow, Jim Wilkinson, the British numerical analyst. Walden: Symmetric matrices and all that. Williams: That’s right—he had worked with Turing at the NPL in Britain. He and I got together and were chatting about Turing, and he was describing Turing and his life in great detail. I can remember dashing up to my room afterward and spending an hour scribbling all this stuff down. Now, of course, it’s all common knowledge; it’s in many biographies; but in those days, it was just not publically well known. Walden: Let’s just run through the rest of your time at Calgary, and then maybe we can go on to your involvement with professional societies, plus the couple of times you had jobs away from Calgary for a period of time; at the Smithsonian, or at the Computer History Museum. So in Calgary, after St. Andrews, you went back and now you’re an Associate Professor . . . then. . . became Assistant Dean for the Faculty of Science. Williams: Yes, and in many ways, that was the most interesting job I had. The Assistant Dean in those days dealt with all the student affairs. You got to see the very, very best students and you got to try to gently guide them along their way because the best students seldom knew what they really wanted to do—they had so many options and choices open to them. And you also got to see the very worst, generally because you had to throw them out. I think I had some teaching relief but that was all. I certainly took the attitude that this is one of the things you did to help out. Nowadays, that attitude seems to be gone from anybody in the department. I think anybody who started the same time I did, or was part of the first cadre, we all chipped in together and helped everybody else out, whereas now, it’s all “well, that’s not in my collective agreement, I’m not going to do it.” That was alien to me. Walden: According to your CV, you spent, what was another sabbatical year, at Chico State? Williams: Yes, and that was because there was a man from Chico State who used to come up to Calgary in the summertime, primarily because it was too hot in Chico in the summer.

April-June 2020

He didn’t like Calgary in the winter. And so he would spend the winters in Chico, and the summers he would come up and teach in Calgary. What I wanted to do was to have a place to go for a sabbatical where they would leave me alone, and take all of these notes and things and try and write some papers, or a book, or something. It’s a small place—it’s noted as a party campus, not a big academic place—but to me, it was an absolute delight. They gave me the biggest office in the place because the chap who was normally occupying it was himself away that year. I immediately got involved in all of the Department’s social activities; they even provided secretarial services for me, which was a real help. I actually wrote my first history book there. Walden: And that was titled what? Williams: It was called A History of Computing Technologyz, and it’s still in print after all these years, though I don’t think it sells very many copies any more. Walden: After your sabbatical, you returned to University of Calgary where, in time, you became a Full Professor. Williams: Yes. It’s hard for me to remember what year these things happened, but there are hoops to jump through and I think probably publishing that book had helped my publishing career get to the point where I was known internationally, because you had to have letters from international people saying “yes, he’s known and wouldn’t bring shame on the university” and things like that. So I’d eventually reached that stage and they made me a Full Professor. Walden: Then somehow in 1986, the Smithsonian called you. Williams: Yes, and that was a big surprise. I now know how it happened because I eventually found out there was a Professor of the history of science at Harvard named I.B. Cohen and I had met him in passing a couple of times. He was interested in computing because Howard Aiken, one of the early pioneers, was at Harvard and Cohen must have read some of the stuff that I did. I went down there for, I think initially, about four months, and then they were looking for a permanent z

Williams, M.R., A History of Computing Technology, Prentice-Hall, 1985. Book is now in its second edition with Wiley-IEEE Computer Society, 1997.

77

Interviews

curator of computers and they asked if I would apply. Now, just at that time, my mother was getting quite old and had developed Alzheimer’s and so she was in a nursing home. Also, my father had started losing his eyesight. I had two siblings left in Calgary, at that point, and they could take part in the family care at the nursing home, but I sort of felt guilty and thought “do I really want to be that far away?” No, I don’t think so. So we made an arrangement that I would leave the Smithsonian but then I would come back in the summertime. And I went back for several summers in a row to help with the curation of the computer project—the exhibit was eventually named The Information Revolution. In the meantime, they appointed an official curator of computers. And I have been quite happy that I didn’t take that job, or didn’t apply for it, because I discovered that the Smithsonian is a very, very interesting place to work with very, very interesting people there, but it’s also rife with politics; it’s worse than the university; it’s worse than almost any place I’ve seen. So I’m glad I never did but I had a very fine time and made some very good friends at the Smithsonian. There was a couple of other sorts of history people who were also involved, so that gave us a chance to collaborate together, as well; one of them being Martin Campbell-Kelly from England. Walden: He was working with the Smithsonian activity? Williams: The same sort of arrangement as me; he’d come for a few months and then go back to the University of Warwick where he was from. Similar things like that. Of course, Paul Ceruzzi was on staff there at the Air and Space Museum and others as well; there were a number of them who made it a lot of fun. I might still be doing that except for one summer, it was brutally hot in Washington D.C. And of course, Washington can be very humid. I was walking into work one morning, because I lived fairly close to the Smithsonian. It was 8:00 a.m. and I was walking through a hot fog; and so I was getting very damp and of course, you then walk into the air-conditioned museum and you start shivering, and at that point I thought, this is crazy. And so I just said okay, I’m just going to quit; I’m not going to come back here. I was sorry to do that, but it was time that I left and let them get on with the job themselves.

78

Williams: Okay. Now at that point, I think I got a call from the Computer History Museum people. There was a gentleman who had been a great benefactor to the history of computing. His name was Erwin Tomash, and Erwin had quite a bit of money behind him. He was one of the founders of the Charles Babbage Institute, which is at the University of Minnesota. He had developed a hobby of collecting books on computing and computingrelated things, and this hobby had sort of gotten out of hand. It got to the point where he had about 3000 books and other items in what he called his old and rare collection and an equivalent number in modern computing from about 1955 on. Now the old and rare stuff was anything from 11th-century manuscripts on up, with a very heavy influence on works from about the years 1500–1700. Some of them were exceeding valuable. At one point—Erwin knew I was interested in books, he had seen me at a conference or meeting of some kind—and he said he was thinking of making a catalog and he needed some help. Would I be willing to help? I said “oh, sure,” because I used to drop by his place in Los Angeles whenever I could, and he and I would play in his library for a fair bit of the visit. At some point, this turned into a much more ambitious project. Now, between Erwin and I, we discussed this for a long time and we agreed that if we’re going to do it, we’re going to do it properly and we came up with a sort of formula where we would have each entry contain the title, the author, obviously, and then the absolutely complete bibliographic information—not just where and when it was published, but even things down to the complete details of the pagination; the collation—which is how the bookbinder put it together—and all sorts of interesting bits and pieces like that. Erwin would do that part and then I would write a text entry, which would start off with something like, “the reason this book is written is . . . ” and if you were talking about Galileo or something. In the end, it took us about 10 or 12 years, but we finished this annotated and illustrated catalog, and it ended up as three volumes, 1,600 printed pages. It is available in print and it’s now also up on the web. It’s hosted by the Charles Babbage Institute, and by the IEEE Computer Society. Amazingly, a few people are actually using it and even the rare book dealers

IEEE Annals of the History of Computing

are starting to quote from it now; not very often but they are starting to quote from it. Well, I was down in Los Angeles working with Erwin. I would take three or four months, drive down and work with him for a while then decide it was time I went home; both of us had gone on like this for a number of years. But while I was down there, I got an e-mail from people I knew at the Computer History Museum in California, in Mountain View, and it said they were thinking of putting a proper building together because at that point it was just a warehouse full of stuff. It was the old Computer Museum from Boston that had moved out to California, stored in a warehouse, and they were interested in a curator for this effort. I took that to mean that they were after me telling them who were the up and coming youngsters who would make a good curator. So I wrote back saying, funny enough, I’m just about to leave here and drive back to Calgary, I’ll just come up that way, stop and we can spend a day together and chit chat. I dutifully showed up in Mountain View, and they said let’s go for supper, because they had a couple of people there from the Board of Trustees, as well. Then I said we should talk about who are the up-and-coming types who you want to think about as a curator. They looked at me and said “oh no, no, no, you misunderstand; we would like you to be the curator.” I did a quick shift of gears and as we talked I started thinking, “this is October, I am driving back to Calgary winter. I don’t like this. This is nice climate down here; perhaps it’s time I actually stayed down here to see what I could do to help out.” To make a long story short, I did drive back to Calgary but then came back down later, because they weren’t ready for me right at that point. They were exceedingly kind. I know of people in the Silicon Valley area who drive for an hour, sometimes even up to two hours to get from home to work; and I just said I’m not going to do that. You’re going to have to help me find a place that I can afford within just a few minutes of the museum, or at least the warehouse, at that point. I told them that I’ve been looking, and if I sold my house in Canada, I couldn’t make the down payment on one close by in Silicon Valley. Therefore, I think maybe with those two problems, I’m not going to come. They immediately came back and said, you can’t afford a house but I can, because the head

April-June 2020

of the trustees board was quite wealthy. If you come down, we’ll go shopping and see if we can find a place that would suit. I figured they’re bending over backward, I have to do this. So I let my brother-in-law live in my house in Calgary and my wife and I just went down there and I became their Head Curator, and basically helped them take it from a warehouse to a proper museum building with about 120 000 square feet. Walden: That’s the building they’re in now? Williams: Yes. And it was just dead lucky we got it because that was the time of the dotcom bust, and Silicon Graphics had recently moved out of the building during the company’s demise. So a few million dollars in modifications later, we had it rehabbed to the point where it became a museum. The next biggest problem was moving everything from the warehouse to the museum, creating storage areas, cataloging everything that we had, because although it started in the Boston museum, by that time, it had grown by about a factor of three. There was everything from tiny microchips to [the 3000 square foot] IBM 7030 “Stretch” computer and multiple Cray supercomputers, all sorts of things. I spent months organizing that collection, or trying to, with lots and lots of volunteer help, and starting on creating the catalog in the process. We got a primitive exhibit space going, which we called Visible Storage. What we did was move things in with enough aisle space so that people could walk around and look, and had a few simple signs in front of things. At that point, I was driving into the museum one Monday morning, and I thought “I have to do this for five more days before I can have a weekend.” I don’t know why that thought occurred to me, because I was enjoying myself there, but I just said okay, you retired once, you can retire again. When I got in I announced that I was going to retire; and part of that decision was that I think by that time I was getting close to becoming President of the IEEE Computer Society; I was working my way up the ladder, anyway. I thought, I can use my time doing volunteer work for the Computer Society; I don’t have to work for a living because I’ve got my university pension. So we parted company very amicably and I still, of course, retain friends down there. Walden: According to your CV, you had begun your move up the ladder at the Computer

79

Interviews

Society, and were moving toward being elected to First Vice President by that time. How and why did you first get involved in professional societies? Williams: It’s sort of interesting; at one time I was a member of quite a number of them. When I was in graduate school in Britain, in Scotland, I was a member of the British Computer Society, and I actually became a Fellow of the British Computer Society. I was a member of the ACM and the Canadian Information Processing Society. Now, I mentioned earlier about this history conference at Los Alamos, the very first one that was ever held. Over supper, sort of an informal gathering, there was talk about lots of different things, but one of them was that we should have a journal. There was a sort of preliminary talk about it; eventually, I think Bernie Galler was the chap who lead the discussion and said “okay, let’s do this”—and he became founding editor of the Annals of the History of Computing. That wasn’t at the Los Alamos meeting, it was a couple of years later. Walden: Which at that time was an AFIPS publication. Williams: Yes. AFIPS was the publisher, because they had the broadest background. Now, at one point, somebody — I can’t remember who it was—probably JAN Lee—contacted me and said are you going to the conference to be held in Las Vegas? It was one of the huge conferences that used to be held twice a year; there was a spring one and a fall one, they called them the National Computing Conferences—the NCC. I said I was thinking of going. He told me that there’s a history session there about Howard Aiken and they need somebody to take some notes and write it up for the Annals. I think at that time he was some lowly scribe for the journal himself, and he said we need to have this write-up, could you do it for us? I told him that I’ve never done anything of that sort but he still asked me to give it a try. So I doodled around; I took a notepad with me and I scribbled and scribbled and scribbled for the afternoon, and I sent it off to the Annals and they published it. Now, when I stop to think about it, that may be where I.B. Cohen first had dealings with me—he’s the one who got me the job at the Smithsonian because he was an expert on Howard Aiken and probably saw my write-up. Now, it’s always been the case that if you’re

80

asked to do something and you actually do it reasonably well and don’t goof it up, they’ll ask you to do something else. You’ve seen this phenomenon before. Walden: Yes, of course. Williams: And so before long, people were asking me to write up something else. At another NCC conference, they were having a side session on the history and that went on several times. Then AFIPS broke up and it was “what to do about the Annals?” I wasn’t involved in anything, except in the most junior way. The editor-in-chief at the time moved the publishing to Springer, who then published it for two or three years. Then Springer did something very strange; it’s probably not strange for a commercial company, but it was strange from the way that we had any dealings before. If you wanted to have a personal subscription to the Annals, it was fairly low in price. If a library wanted to have it, it was quite high in price. Most of them [academic subsciberrs] immediately canceled their subscription, so the Annals was suddenly in trouble, and Springer didn’t seem to care. So JAN Lee had been interested in, and involved with, the IEEE Computer Society before, and he suggested that perhaps the Computer Society would take over the publishing of the Annals and [he] made a very passionate plea to them. And they eventually said yes, they would do it. So suddenly, Annals became part of the publishing of the IEEE Computer Society. Now, at that point JAN Lee took over as the Editor-inChief and he phoned me up and he said, “we have a problem, we haven’t got anything to publish.” He said he had gotten stuff from Springer, there were no papers in the pipeline at all, and they had—whether in retaliation or what, I don’t know—destroyed all the back issues of the Annals that were in their possession; stuff that had been shipped to them from AFIPS. So you couldn’t buy any back issues of the Annals at all, so there were very, very few complete runs of the Annals anywhere. And he said, “what are we going to do?” Incidentally, he also said, “I want you to be my assistant Editor-in-Chief.” I suggested that we might start by him writing a paper; I’ll write a paper; we will find something we can publish and basically we will fill out an issue. JAN said that’s not good enough because we’re already

IEEE Annals of the History of Computing

late for one and the next issue is coming up. So I said “good, we’ll make it a combined issue,” and JAN and I sat down and tried our best to write something, a research article each. I can’t remember who found it, either JAN or I, but we found stuff that von Neumann had done, and we published that. We also later discovered that it previously had been published in the Annals and we hadn’t recognized it. But, you know, somehow or other we got it limping along and JAN said “look, it’s time you came to an IEEE Computer Society meeting with me to see what’s going on.” Basically, I just listened to JAN and then we went off and did something else. That sort of got me sucked in and when JAN had finished two terms as Editor-in-Chief of the Annals—by IEEE rules, he could only do a maximum of two terms then he had to step down—I said ‘okay, how do I apply for the job?’ I filled out no end of stuff and sent it into the IEEE, and I never heard a thing. Nothing! Months went by, never heard a thing. Walden: And when you say IEEE, you really mean sending it to the Computer Society? Williams: Yes. At one point, I phoned up JAN and said, so who did get to be editor? I got phoned up and it’s probably because I asked JAN Lee what was going on, and he passed it further on; said somebody better tell him he’s the editor before it goes too far. I had no intention of doing anything other than the Annals, I was a historian by this time and that’s all I was interested in. But it turns out that after doing this for four years, that was two terms, there were a lot of very interesting people I met and at one point, my term was coming to an end and I had done as much as I could to prepare other people to take over. But it was about the last meeting I was in, and there was always a hospitality suite at these meetings; so I was up at the hospitality suite after all our business was over, and there was a woman there. I had no idea who she was; sitting down beside here and making idle conversation. I said you know, I really enjoyed this time as EIC of the Annals. How do you keep doing other things with the Society? She said oh, you want to do more work? I said that I thought that might be fun. She said, “Well, just leave it;” so that was the end of the conversation. I had no idea that I was talking to the person who was the head of the nominations committee for other jobs. I honestly had no idea at all, but before long, I was asked to do a

April-June 2020

couple of other things and then, as I said before, if you do even a half-heartedly good job they’ll ask you to do something else. And after, let me see, I was Editor-in-Chief of the book publishing program for a while; and then they asked me to chair other publishing-related committees, and eventually I was appointed as the Vice President of Publications, because that’s a Presidential appointment, it’s not an election. The Publisher at the time was Angela Burgess and Angela and I got along quite well, and probably when it came time for the President to find somebody, Angela probably said well, there’s Mike. And so I got in that way. But somehow I became the Vice President, Publications, and the first meeting, they had to find somebody to go to the IEEE level and represent the Computer Society in the IEEE Publications Board. Traditionally, if you’re Vice President of Publications of the Computer Society, you represented the Society at the IEEE level Publications Board. Now, it turns out that there was somebody else who would have liked to have done it, Jerry Engel, who was becoming a quite good friend of mine, he was the President of the Society himself; and he and I had to go outside the room while the Board of Governors had a long chat about us both. And again, it was the thing about if you seem to be doing something reasonably well, they’ll ask you to do something else. Whether you’re qualified to do it or not doesn’t seem to matter. And so I moved around a fair bit, not only with the IEEE Publications Board but with their Regional Activities Board, later called the MGA, the Members and Geographic Activities Board. And as you know, I eventually, actually won election to the Presidency of the Computer Society. Walden: You were competing against a fellow named Yervant Zorian? Williams: Yes. And I was told afterward that I was the sacrificial lamb because I was an academic and Yervant Zorian was the practicing engineer, and practicing engineers always won. I was meant to be the sacrificial lamb and nobody was more surprised than me and Yervant when I won and not him. Walden: Well it’s a very interesting pair of [Personal] Statements. He speaks about the good of Society—“I’ll serve the members,”— quite a traditional presidential statement. You

81

Interviews

speak about basically “the society’s in trouble and we need to do a lot to fix it,” and presumably some of that came because you’re interfacing with the whole IEEE, you’re involved with publications, you know Angela, and so on; and I’m wondering to what extent the members somehow thought, “oh, we’ve got somebody who thinks it’s in trouble and it needs to be fixed. Maybe we should get that guy rather than the guy who’s saying let’s keep doing good services.” Williams: It’s entirely possible. I mean, everybody who casts a vote probably has a different rationale and, of course, there are only a very small percentage of members who actually cast a vote. Walden: When I read this, it’s quite a startling set of contrasts between the two statements. Williams: I suppose it is. Walden: And then your first presidential letter in Computer is “this is the year we have to decide things.” Williams: Yes, that was; the Computer Society was starting to lose money. The year before, we had run a considerable deficit and before, there was really no deficit ever. That may not be exactly true but the finances were always in good shape. We had a fair bit of money in reserve but we still ran $500,000 or something in deficit. This is not my style; I don’t like deficits. I used to joke with people because the president has a great deal to say about where meetings were held, and there’s always about three meetings a year. Now, in those days, you could basically specify where those meetings are held and my predecessor had held some in Hawaii, and there were people who held them in all sorts of interesting places. I used to joke that I was going to hold them at the Motel 6 in Pierre, South Dakota and this actually became a running joke. A number of people have mentioned it to me recently, although they never think it’s Pierre, they think it’s someplace else. However, it was my idea that it was time to tighten our belts and stop spending. Angela Burgess was also of this ilk. Now, we had gone through a tough time earlier. There was at one time an Executive Director of the Computer Society, back about the year 2000, named Michael Elliott—a no-nonsense individual who liked being the Director of a large organization. He had worked for Bill Clinton in

82

Alabama; he was on Bill Clinton’s cabinet there, I think, in some respect. He basically ran the Computer Society and he ran it as an independent entity. Nothing was ever referenced to the IEEE as a whole, or seldom anyway. All the decisions were taken internally and had been done that way for quite a number of years. The IEEE needed a new executive director themselves. When Angela Burgess finally became Executive Director, she said okay . . . “we’re missing out on a lot of stuff by not cooperating with IEEE.” And that is the same time that I was President; because it was in my watch that we had to hire Angela. . . . We will even make one of our yearly meetings jointly with IEEE so that they can see what we’re doing and everybody can meet everybody else. If you get to know the other people, it’s often a lot easier; and it was the fact that we had also lost half a million dollars the year before. Now I came out the end of my year as President with a surplus. Not much, it was $50,000 or something, but it was a surplus. And I think for several years thereafter, there was nothing but deficit after deficit after deficit. I don’t know what the situation is today, but the membership is falling; the income from traditional journal sales is falling, but that’s universal. They’re doing their best to fight it off. There were a couple of us who thought Angela was an exceedingly good person to guide and to actually run the Computer Society. She had just recently completed an MBA and she probably knew more about how the IEEE and the Computer Society worked than anybody else. We had a job search in which the legwork was done at the IEEE level by their staff; and there were about four or five people on a short list including Angela, some of whom were very, very good. It came down to three, and the selection committee had to meet in Washington, D.C., and we were going to have supper with each of these people, just to see if they could use the right fork, at the same time as well as being decent administrators, and, you know, didn’t blot their copybook in any way. But we didn’t have time to meet with Angela, and so we scheduled a lunch with Angela. Now, Angela said afterward, that as soon as we did that, she realized that this was pro forma, that she was not to have the job because if she was going to have the job we would’ve done supper

IEEE Annals of the History of Computing

with her like we were doing with everybody else, at a very fancy hotel, instead of a quick lunch in a not all that fancy a place. It was also a long weekend; probably President’s Day or something; I can’t remember. She said, you know, had they been polite and said look, you’re not really in the running so we won’t bring you out here; enjoy the long weekend with your family. She said she would’ve actually appreciated that and thought that we were all rascals because we didn’t do it that way. And so when we all got together after this and had a vote, and Angela was the clear winner, we got a speakerphone in the middle of the table and phoned her up and said, Angela, how would you like to be Executive Director? And she almost fell apart because she was absolutely so certain that she hadn’t got the job. All the signs were wrong. But it’s just that she was so good that I don’t think that we had to worry about it, except perhaps for one person on the selection committee who didn’t really know her and did their best to derail her selection. Interesting, but they were outvoted. Walden: I’m curious about how you see the whole operation functioning. There’s the President Elect, the President, and then the Past President, and you’re only President for a year, and the staff is basically running operations. Williams: That’s right. Walden: . . . it seems hard to get things done in, for instance, your first letter in Computer, you talked about how Carl Chang, several years before had said there needs to be transformation. . . How does all that actually work in terms of having some consistency in getting things done? Williams: It’s very difficult. And basically, the president is more of a figurehead than anything else. As you said, the staff runs the show. Now, technically, the President does have some powers. It was always my view that the Executive Director deals with staff matters. A few presidents in the past had tried to intervene when somebody was being fired or was threatened to be fired, or something and this often lead to disaster. The Presidents and Vice Presidents certainly have influence on the strategy and the direction, and if the Board of Governors says we will do this, then the staff will simply do it. Walden: So they then have to figure out how to implement it?

April-June 2020

Williams: Yes, and sometimes it’s not easy. . . but usually there were very good relations between staff and the executive and the volunteers. There was one problem when I was President, which I want to tell you about just to get it on record. I, of course, had things that I wanted to do, and one of the things that basically had been quietly arranged amongst the various staff was we had this beautiful old building on Massachusetts Avenue and Embassy Row in Washington, D.C., and IEEE U.S.A., the American version of the IEEE, was headquartered in Washington and so were we. Our publishing was done in California but technically the headquarters was Washington. They had sort of agreed amongst themselves (because the IEEE was having to renew a lease) [that] perhaps IEEE-USA should move in with us. We owned the building, they could pay rent, and everybody would be happy. We would have to do some renovations to that building and so the engineers were called in to see about the renovations. There was not only the big building but there was an office complex at the back that nobody ever saw, like a sort of coach house or something, with a bunch of offices and things in it and we weren’t exactly sure how to modify this. At one point, my phone rang and it was the head staff financial man at the IEEE who asked if I was going to a meeting to be, I think, held in Portland next week. I walked into the meeting and I could see him there so I went over, tapped him on the shoulder and asked “what’s wrong?” We went out in the hall so we got a nice quiet place to talk and that’s when he said that he got the report from the engineers and they had said that the electrical system was shot and there was going to be a fire before long. And that is, if the carbon monoxide from the heating system didn’t kill anybody first. In fact, they had condemned the building and we had seven days to get everybody out. And that basically. . . made all my plans come to nothing because the rest of my presidency was circling around that and all of the chaos that came from it. Now, it turned out there were holes in the chimney and things, holes in the furnace, and at some stage in the past, somebody had decided they had to put lightning rods on the roof, and had run the grounding wire down behind the fire escape at the back. It was a metal fire escape on the back

83

Interviews

of the building and to do that, they had unscrewed the fire escape and had run this line but then they had forgotten to screw it back on again so that had there been a fire and had people tried to leave on the fire escape it would’ve collapsed under all the weight. It was just one thing after another, after another. There were also fire escape routes that were blocked. It turns out that we had to either basically tear it down and rebuild or spend a very large amount of money in an attempt to rehabilitate it. If we tore it down, Washington, DC regulations said we had to keep the front fac¸ade. The Executive Director of IEEE U.S.A. and I went around to see the architects and we had several meetings with them about what could be done and roughly how much it was going to cost. I can remember it was between $7 and $9 million. Eventually, it came down to having to make the decision; I wanted to actually stay there because it was a nice building, nice address, and everything else. The rest of the Board of Governors decided that we shouldn’t be in the real estate business, that wasn’t our job and they voted to sell the building, which I accepted. That was a reasonable decision. It didn’t go the way I wanted it to but then we didn’t have a spare $9 million on hand so this was perhaps the most reasonable decision. Walden: Where had the staff gone before the week was up? Williams: They had tried to move in with IEEE U.S.A. So instead of them moving in with us, we moved in with them and we certainly didn’t have enough room for both organizations. We were stacked in hallways, and we were sharing offices. . . They were very good about it, they couldn’t have been nicer, and they really saved our bacon. But then followed no end of trips around Washington, looking at real estate. There were buildings that were being built and we could move into, and there were other buildings that were already there that were being renovated that we could move into. I can remember Angela Burgess and I, and a few others, including people from IEEE, traipsing around Washington D.C. just looking for someplace and finally, I think it was Angela and the IEEE U.S.A. who said “this is the best one.” I sort of said you have to live in it; you go ahead and make the decision.

84

Walden: So IEEE U.S.A. went to the new building, too? Williams: Yes. Walden: So you did share a building, just not the building you intended. Williams: We shared a floor on a bigger office building. It wasn’t the building we intended, but everything worked out in the end. That took a huge amount of my presidency; took my eye off of no end of other goals. Walden: During this time, if I read your reports to the Board of Governors or your informal newsletters that you sent out, which I have, you’re also struggling with the IEEE and Computer Society financial treaty. Was that your doing that or was Angela mostly handling that? Walden: It was mostly Angela and the finance people on either side. We actually had several rules, and one of them was that if you have or if you run a deficit, and can’t remember the exact rule—but I think it’s two years in a row or something three years in a row perhaps—then you’re put on what’s called a “watch list.” While everybody is, or at least was at that time, running around saying oh, we can’t go on the watch list that would be terrible; it turns out to be a paper tiger. And we have been on the watch list, I think, ever since because we hadn’t run enough positives in the budget. Walden: So you cease to be the President and became the Past President. Did the new President continue sorting all this out? Williams: Yes. The new president was Rangachar Kasturi. He is a good guy; I have great respect for him. He is a Professor in Tampa, Florida, and I think he’s one of the smartest men I’ve ever come across and the hardest working. He is so workaholic that at one point, he and Angela Burgess used to compete as to who would send e-mails at 2:00 in the morning to one another. Angela had to finally give up because she was so exhausted she was going to end up in the hospital, or worse. Rangachar just kept doing it. . . . But to get back to your question, as President Elect, you do a fair amount of stuff. A lot of this is internal chairing of miscellaneous committees; you chair this appointments committee; you’ve gotta read the constitution, but you chair an awful lot of stuff. And when you become president . . .

IEEE Annals of the History of Computing

. . . it’s very often the case that you spend an awful lot of time going around shaking peoples’ hands, so you’re often more of a figurehead than anything else. My wife kept track, and I was away from home 157 nights in that year, so about half the time. An awful lot of that was going to some conference to present an award to somebody or other, and so you leave, say, today to fly someplace; the next day to attend the conference, present the award; and the next morning you fly home again; you’d have one day at home; and the next day you fly someplace else again. There was a lot of that kind of stuff. For example, there was a computer pioneer named John Vincent Atanasoff, who was born in the US but his parents came from Bulgaria. The Bulgarians are very proud of him and have issued stamps, and all sorts of things with the fact that he’s a computer pioneer [of theirs]. His family tried to make an argument that he was the inventor of the computer—legally, in the US, that is correct because it was the result of a verdict in a patent lawsuit, but the needs of patent lawyers are not the same as the needs of a historian, so you’ll find historians will often say that it’s not quite correct. But Carl Chang—Computer Society President in 2004—is from Iowa State University and that’s where Atanasoff worked . . . at least for part of his life. Because of that, Carl Chang was interested in trying to get an Atanasoff Prize established by the IEEE. You have to give about a million dollars to the IEEE to set this Prize up and Carl, with a bunch of other people who are well connected in Bulgaria, got the Bulgarian government to agree to fund a million dollars. Now, getting them to agree to fund a million dollars and actually getting the million dollars are two different things. It went for several years with “oh, well, we’ve got to get it approved by so-and-so; and yes; in the meantime, it’s in somebody’s bank account.” Eventually, I was part of several delegations that Carl Chang had going to Bulgaria. There was a big conference in Sofia, it was to celebrate Atanasoff; there were arrangements made to meet the Bulgarian President, and things like that. And that was during my presidential year and so I was along there, and making speeches at the conference as well as shaking hands with the President who didn’t care who any of us was. He wanted to get out of

April-June 2020

there as fast as he could. I don’t think that money has arrived to this day, and I don’t imagine it ever will. There’s a lot of that kind of stuff that you do as President. As Past President, there is also a lot of sort of internal stuff, and it’s all laid out in the constitution; what you chair, and when you do this, and when you do that; but there was also a few things where you had to go up and shake hands again. Walden: As the president, are you in touch with the Executive Director daily? Weekly? Williams: Oh, pretty much daily or every second day. There’s occasional times when we agree to have a down period for a couple of weeks—“I think we’re doing alright here”; or “there’s no actual crisis at the moment.” But I was certainly in touch every few days with the Executive Director, either the one preceding Angela or Angela. The one preceding Angela was a chap named David Hennage, who was a very pleasant man. Walden: I’d like to move on from your IEEE and Computer Society activities. Before I do that, I’d like to mention that . . . you did continue with the IEEE, with the Computer Society in various committee roles; as chair of the IEEE History Committee, as a member of the Computer Society History Committee, and so on. Williams: Yes. In fact, I chaired the IEEE History Committee twice in my career; once in the 1990s and once in the 2000s; I forget whether it was 2009, or something. So I’ve been involved with them for a lot of times. Walden: In the early days, I think, of computing history, a lot of people came from the world of computers and sort of drifted into history. It’s your case, it’s Martin Campbell-Kelly’s case, it’s the case for others. Increasingly these days historians of computing science study history. Williams: Yes. Walden: . . .And are not out of the computer world, although may have an undergraduate degree in computer science, but perhaps they never really practiced it. Williams: Right. Walden: And the other question is, the Annals, it’s clearly increasingly dominated by what I’ll call professionally trained historians. Williams: Yes, very much so.

85

Interviews

Walden: What’s your view of the historians who didn’t come from the practitioner role, either the teaching world of computer science or the engineering world of computer science sort of dominating this field, now? Williams: It certainly changes the whole complexion of the field. The early engineers and computer scientists were more concerned with the internal history, as it’s called. Who built what? How did they build it? What equipment did they use? What vacuum tubes? All of the minutiae of detailed technical stuff. The modern trained historians coming out of the history of science program usually do not understand any of the technical side and because of that, they are not interested, and what they are interested in is the social impact that computers have had. Now, I think that particular aspect, obviously, the computer has had a huge social impact, but I don’t much care. I’m more an internal historian. Okay, so it really is a different complexion to everything and occasionally, you find the engineer making completely ridiculous statements about societal effects and you find the social historian making completely ridiculous statements about technology. One thing I might mention: about five years ago, I got a missive from a chap that I know quite well, who is on the ACM History Committee. The Committee had decided that it was time to get a new version of the Turing Award website with more up-to-date information about the Award winners. Now, for some of them, there was a page of information there; others there was just, you know, “Joe Blog won in 1970-something” and that was all; not why he won it, nothing about

himself. So it varied from reasonable stuff to just completely ridiculous, and would I be willing to take part in this? I said no, I still have stuff to do with the Computer Society and I don’t want to bite off more than I can chew. They said “oh, we understand” and it was a couple of years later that I got this phone call again from this same fellow who said I know we asked you once and you said no. But, he said, you’re not quite so busy anymore and we just got approval from the ACM to do this, and we’ve got approval to pay the writers, but what we need is somebody to sort of oversee this thing and make sure that it’s all done properly. Would you be the overseer? And so I finally said yes; and now, if you look on the ACM site for the Turing Award winners{ you’ll find that every person has a large write-up. A few were done by me because I couldn’t find anybody else who would write it so I just did it myself. Others were done by professional writers; others were done by former colleagues, etcetera. Did you do one? Walden: I did the one on Knuth. Williams: I thought so. And that is now finished, and I’m out of it. I understand that Tom Haigh has agreed to do the oversight stuff; and now, of course, you only have to keep it up-todate every year, it’s no big job. When you have to do 70 or 80 it’s a much larger effort. But it looks good, and I’m quite proud of the fact that I was there to at least guide it along and then find people like yourself to move it along. Walden: Wonderful. Okay, well thank you very much, Mike. You’re having a fascinating life and I’ve really enjoyed hearing about it. Thank you very much for coming. Williams: Thank you, Dave, it’s been a pleasure.

{

86

https://amturing.acm.org/byyear.cfm

IEEE Annals of the History of Computing

Book Review

Meryl Alper, Giving Voice: Mobile Communication, Disability, and Inequality. Cambridge, MA, USA: MIT Press, 2017 Nabeel Siddiqui Susquehanna University

& IN

HER WORK, Giving Voice: Mobile Communication, Disability, and Inequality, Meryl Alper explicates “the social implications of communication technologies that purport to “give voice to the voiceless” by foregrounding issues of disability and access (2). Hyperbole surrounding augmented and assistive communication (AAC) devices professes that they serve as tools of liberation for individuals with communication disabilities. In contrast, Alper shows how race, gender, and class accentuate inequalities amongst those with disabilities utilizing AAC devices and how policy makers, manufacturers, and parents rarely acknowledge these misgivings. By exploring the ways children with communication disabilities incorporate AAC devices into their everyday lives, Alper complicates the techno-utopian narratives surrounding assistive technologies and

Digital Object Identifier 10.1109/MAHC.2020.2991702 Date of current version 29 May 2020.

April-June 2020

provides an intervention for those who may see disability through a model emphasizing deficit and lack. Throughout her work, Alper draws on a myriad of sources and approaches to explicate “voice,” but her primary data stems from participant observation and in-depth interviews amongst families of children with communication disabilities. The families span a range of class, gender, and racial categories but are united by their use of iPads with the Proloquo2Go application. Created by AssistiveWare, Proloquo2Go transforms the iPad into a “symbol based” AAC device that purports to be a “voice for those who cannot speak.” For the author, this claim places the application in a broader, more problematic rhetorical framework surrounding disability. As she argues, “voice is an overused and imprecise metaphor—one that abstracts, obscures, and oversimplifies the human experience of disability” (5). Consequently, she implores scholars to address the

Published by the IEEE Computer Society

1058-6180 ß 2020 IEEE

87

Book Review

extensive social, cultural, and political factors that entwine with assistive technologies. Giving Voice foregrounds the personal accounts of those using mobile communication technologies to illustrate inequality and can be separated into three sections. In the first, Alper looks at “voice” as a rhetorical tool that surrounds the history of AAC devices. As she makes clear, the metaphor of “voice” obstructs a more refined view of how structural privilege, especially class, shapes the use and reception of AAC devices, and parents’ beliefs of their children’s voices often differ from the dominant discourse. In the second section, Alper explores the iPad as a material entity. She finds that parents imbued not only the iPad but also its protective cases with specific values. These values diverged based on identity markers and reflected desires to distinguish between multiple devices. In her final section, she assesses the relationship between youth and media. In contrast to popular conceptions of mass media as detrimental, parents of children with communication disabilities saw media as reflective of a cognitive sophistication neglected by outsiders. At the same time, they held a critical view of media representations that perpetuated stereotypes of disability, which they hoped to counter. In overviewing first-hand accounts of those with communication disabilities and their families, Giving Voice shows that assistive technologies can take on a range of meanings, but the most engaging aspects of Alper’s work is her analysis of how mobile communication devices challenge our neurotypical and able-bodied assumptions of everyday life. As the first book length study of the iPad and its adoption, the author provides a road map for scholars interested in the intersection between inequality and mobile technology, and shows that the ubiquity of mobile communication devices presents its own set of challenges as issues of race, gender, and class are often embedded into technological artifacts themselves. Furthermore, because these identity markers intersect with the culture of disability, Alper makes apparent the need to counter the dominant discourses surrounding assistive technologies and for scholarship that moves beyond technological adulation. Rather than focusing purely on critique, Alper strengthens the credibility of her research by providing recommendations for policy makers,

88

technology developers, the mass media, and practitioners hoping to address inequalities perpetuated by assistive technologies, and it is here that her work most shines. She argues for a holistic approach to evaluating AAC devices that takes structural inequality seriously rather than defaulting to a naive belief that technology provides “voice.” She asks us to inspect our own implicit biases and urges technology developers to take seriously the difference amongst those with disabilities. Finally, she implores policy makers to look at students on a case by case bases and for the mass media to showcase, “a broader array of backgrounds, not just white, uppermiddle-class individuals whose visibility is already heightened by the access to social, cultural, and economic resources.” If there is one chief flaw of Giving Voice, it is that its foregrounding of class results in only a brief overview of race and gender. When Alper does discuss these concerns, she focus more on the technological design of the Proloeue2Go rather than the entanglement of disability with the broader cultural, social, and economic frameworks surrounding it. For instance, in her analysis of the available synthetic speech options of Proque2Go, Alper laments the limited choices and diversity in the software. As scholars such as Rosemarie Garland-Thompson, Alison Kafter, and Robert McRuer show, however, to truly understand disability, we need to take an intersectional approach to disability itself and not just its manifestation in technology. Although Alper notes this broader literature on disability and intersectionality, in practice, its implication become bypassed for a detailed focus on class. Nonetheless, Alper’s work is an impressive analysis of mobile communication device and disability. Her complication of “voice” showcases the complexity of discourses surrounding AAC devices, and her engagement with families of children with communication disabilities provides a rare purview for scholars and readers interested in exploring these devices in more depth. Perhaps most importantly, Giving Voice serves as a paradigm shift for digital media research and shows the necessity for more exploration of how issues of structural inequality shape and have shaped the development, reception, and use of assistive technologies.

IEEE Annals of the History of Computing

HOST

ER T IS G RE

2020

W O N

!

6–9 Dec. 2020 • San Jose, CA

IEEE INTERNATIONAL SYMPOSIUM ON HARDWARE-ORIENTED SECURITY AND TRUST 6–9 Dec. 2020 • San Jose, CA, USA • DoubleTree by Hilton Join dedicated professionals at the IEEE International Symposium on Hardware Oriented Security and Trust (HOST) for an in-depth look into hardware-based security research and development. Key Topics: • Semiconductor design, test  and failure analysis

• Cryptography and cryptanalysis

• Computer architecture

• Imaging and microscopy

• Systems security Discover innovations from outside your sphere of influence at HOST. Learn about new research that is critical to your future projects. Meet face-to-face with researchers and experts for inspiration, solutions, and practical ideas you can put to use immediately.

REGISTER NOW: www.hostsymposium.org