Computer Applications and Quantitative Methods in Archaeology 1989 9780860546955, 9781407348438

Table of contents :
Front Cover
Copyright
Contents
1. Preface
2. Investment appraisal for information technology —the experience of the English Heritage record of scheduled monuments
3. Computerised county sites and monuments records in England —an overview of their structure, development and progress
4. SUPERFILE—a user's view
5. Using dBase for county SMRs: the Humberside experience
6. The Greater London Sites and Monuments Record —a case study
7. Sites and Monuments records in historic towns —problems in the development of computer records for urban areas
8. Pooling Resources: drawing on regional data from the computerized Sites and Monuments Records of the East Midlands
9. The National Archaeological Record
10. The ArcheoDATA Project
11. Algorithms for the enhancement and reconstruction of archaeological data sets
12. Image Processing In Archaeological Remote Sensing
13. Rectification of aerial photographs by means of desk-top systems
14. Image segmentation techniques for archaeological geochemical data
15. Recognising and controlling for cultivation-induced patterning in surface artefact distributions
16. Mathrafal: a case study in the application of computer graphics
17. Applying Solid Modelling and Animated Three-Dimensional Graphics
18. Deconvolution of ^T profile curves
19. Methods for Processing Digital Geophysical Data in Archaeology
20. Fifteen years of contributions of the French school of data analysis to quantitative archaeology
21. Seriation and Multivariate Statistics
22. Seriation by similarity and consistency
23. Principal Components Analysis of compositional data in archaeology
24. The chemical composition of tiles from Bordesley: a case study in data treatment
25. Geodetic and cartographic problems in archaeological data bases at and within the boundaries of some countries
26. Error structures of ceramic assemblages
27. The development of an integrated archaeological software system
28. The COMPASS System: Computer-Assisted Surveying and Mapping for Archaeological Fieldwork
29. Archaeological computing in South- Western Europe (France, Spain, Portugal and Andorra)
30. Hypercard as a teaching tool
31. Beyond classification: the use of artificial intelligence techniques for the interpretation of archaeological data
32. Shape Information in an artefact Database
33. The analysis of ancient Chinese pottery and porcelain shapes: a study of classical profiles from the Yangshao culture to the Qing dynasty using computerised profile data reduction, cluster analysis and fuzzy boundary discrimination
34. Rule-based Dating of Artefacts

Citation preview

Computer Applications and Quantitative Methods in Archaeology 1989 edited b y

Sebastian Rahtz Julian Richards with Ben Booth . Clive Orton Gary Lock Clive Ruggles

BAR International Series 548 1989

B.A.R.

5, Centremead, Osney Mead, Oxford OX2 ODQ, England.

GENERAL EDITORS

A.R. Hands, B.Sc., M.A., D.Phil. D.R. Walker, M.A.

BAR S548, 1989: 'Computer Appli.cati om..; and Qudnlitative Xt?thod:.:i in Archaeology, 1989 1

© The lndi.vidual Authors, 1989 The authors’ moral rights under the 1988 UK Copyright, Designs and Patents Act are hereby expressly asserted. All rights reserved. No part of this work may be copied, reproduced, stored, sold, distributed, scanned, saved in any form of digital format or transmitted in any form digitally, without the written permission of the Publisher. ISBN 9780860546955 paperback ISBN 9781407348438 e-book DOI https://doi.org/10.30861/9780860546955 A catalogue record for this book is available from the British Library This book is available at www.barpublishing.com

Contents 1

vii

Preface Julian Richards

Sebastian Rahtz

I 2

CULTURAL RESOURCE MANAGEMENT Investment

appraisal

for information

edited by B. K. W. Booth technology

1 1

Nigel Clubb

2.1 2.2 2.3 2.4 2. 5 2.6 2. 7 2.8 2.9 2.10 2.11 3

Introduction.... ................. .................... ......... ............................................... Investment appraisal as defined by HM Treasury..................................... The context of investment appraisal............ ............................................. Investment appraisal and 'willingness to pay'....... ......................... .......... The English Heritage record of scheduled monuments.. ......................... Net Present Value..................... ................. ............... ............ ........................ Quantifiable costs and benefits.. ......................... ...................... ................. Alternative solutions..... ................... .............. ............................................. Unquantifiable benefits ................................. .............................................. Uncertainties...... ....................................... ................... ................ ............... . Conclusions............................... ..................................................................

Computerised

county sites and monuments

records

in England

1 1 2 2 3 4 4 4 6 6 7 9

Amanda Chadburn

3.1 3.2 3.3 3.4 3.5 3.6 3. 7 3.8 4

Introduction ....................... ............................. ............. ................................ The history of SMRs .................................................................................... The scope and content of SMRs................................. ................................. Data structure................. .................................................... ......................... Back-up material....... ............. ...................... .................... ............................ Computerisation......................................................................................... . Compilation to date............................... ...................................................... Conclusions.................................................. ..................................... ..........

SUPERFILE-a P. D. Iles

4.1 4.2 4.3 4.4 4.5 5

user's view

9 9 10 10 12 13 17 17 19

M. Trueman

History of the Lancashire SMR............. .... .... ...... .. .... .. ........ ........ .. .... ........... Superfile ................ .......... ............................................................................ Problems ................. ......... ............ ................. .......................................... .... An easy solution? .......................................... .......... ................................... Conclusion .......................... ... .............. .......... .............................................

Using dBase for county SMRs: the Humberside

experience

19 19 20 22 26 27

John Wood

5.1 5.2 5.3 5.4 5.5

Introduction................................. ................. .................. ............................. Why create an SMR?.................................................... ............ ...... ........... .... Who will use it?......................... .................... .......... .................... ................ The strategy adopted .................. ............................. ............. ............ .......... dBase: Why and How...... ................ .................. ...........................................

27 27 28 28 29

CONTENTS

6

The Greater

London Sites and Monuments

Record-a

case study

33

Hugh Jones

6.1 6.2 6.3 6.4 6.5 6.6 6. 7 6.8

Introduction.............. .......... ............................ ......... .......... .......................... History of the SMR.................... ................................... ..... ............. .............. Conservation in London.............................................................................. The structure of the record...................... ......................... ......................... The contents of the record........................ ................................. ................ The current computer service.................... ............................. ................... Hardware and software. .............................................................................. Computing blind alleys............................................................................... 6.9 Advantages and disadvantages of the current computer system....... ..... 6.10 T he future direction of Greater London SMR computing.................... ..... 6.11 Mapping..................... ........................ ........... ................ ......... ....................... 7

Sites and Monuments

records

in historic

towns

33 33 34 34 35 35 36 37 37

38 38 41

N. A. R. Lang

7.1 7.2 7.3

8

Introduction................................................................................................. Inquiries to SMRs...................................................... ................................... Database managers and data -content.......................... ......... .................. ...

Sites and Monuments Roger Martlew

8.1 8.2

8.3 8.4 9

John Hart

10

43

51

Barny Creaser

Introduction and background... .......................... ................................. ....... Leicester University Regional SMR Database ............................................. The East Midlands SMR system: a test application.... ............................ ... Conclusions .............. .... .... ........ .. .. ........ .. .. .. .. .. .. .... ........ .... ...... .... .... .. ...... .. ...

The National

9.1

Records of the East ·Midlands

41 43

Archaeological

Record

51 52

54 54 57

Roger Leech

Introduction......

The ArcheoDATA

............................ ..................................... .................... ...... Project

57 69

Di1niel Arroyo-Bishop

10.l : 0 .2 10.3 10.4 10.5

Introduction ................................................................................................ The ArcheoDATA Project ............................................................................ Preparing the archive ................................................................................. The material implementation ........................... ......................................... Conclusion ............................................................................... ...................

II SURFACE AND SOLID MODELLING AND IMAGE ENHANCEMENT ARCHAEOLOGICAL DATA edited by C. L. N. Ruggles 11

Enhancement

and reconstruction

of archaeological

69

70 77 80 86

FOR

data sets

87 87

]. E. Szymanski

11.1 11.2 11.3 11.4 11.5 12

Introduction....... ....................... .................... ....... .................... .................... 'Classical' data processing............ ................... .................................. ......... Simulated Annealing.... ............................................................................... Maximum Entropy....................................................................................... Overview......................................................................................................

Image Processing ]. G. B. Haigh

In Archaeological

Remote Sensing

87 88 89

90 93 99

S. S. Ipson

12.1 Introduction............................................................. ii

....................................

99

CONTENTS 12.2 12.3 12.4 12. 5 12 .6 12. 7

The image processing system.................................................................... System software .......................................................................................... Application to an aerial photograph.......................................................... Application to a resistance survey............................................................. Implementation on a PC.............................................................................. Conclusions .............. .. .. .. .......... .. .. .. ...... .. .... .. .. .. .. .. .. .. .. .. .... .. .. .. .. .. .. .. .. .. .... .. .. .

13 Rectification

of aerial photographs

by means of desk-top

systems

99 101 101 103 106 108 111

]. G. B. Haigh

13 .1 13.2 13.3 13.4 13.5

Introduction................................................................................................. The use of the projective transformation................................................. Introduction of photogrammetric techniques........................................... Timings and portability..................................................................... ......... Current state and future prospects ............ ................................................

14 Image segmentation 121 C. E. Buck

14 .1 14.2 14.3 14.4 14.5 14.6 14. 7

techniques

for archaeological

geochemical

data

C. D. Litton

Introduction........................................................ ......... ................. ............... General methodology................... ....................... ..... .... ......... ...................... Detailed methodology ..................... ............................................................ Simulation .......................... ........... ......... ........... ..................................... ...... Example ................... ........ ............................................ ................................. Discussion ....·................................................................................................ Acknowledgements ............... ......................... ...................... .......... .............

15 Patterning

111 111 114 116 118

in surface artefact distributions

121 122 123 125 128 128 130 133

W. A. Boismier

15.1 15.2 15 .3 15 .4 15. 5 15 .6

Introduction ................................................................................................. The dataset ................................................................................................. Changes in artefact frequency...................... ............. ................................. Horizontal displacement ............................................. .................... ........... Minimising horizontal displacement patterns.... ......................... ............. Conclusion................. .... ..................... ........... ................. .............. .... ...... .....

16 Mathrafal: C.]. Arnold

16 .1 16.2 16.3 16.4 16.5 16.6 16.7 16.8

a case study in the application ]. W Huggett

P. Reilly

of computer

graphics

147

C. Springham

Introduction.................................................................. ............................... The site ........................................................... :............................................. The Mathrafal Project.................................................................................. The application of computer graphics........ .................................. ......... .... Methodology.............................. ..... ............................................................. Computer analysis of the Mathrafal survey data.......................... ............ Visually linking the collected data and interpretations ........................... Conclusions... ........ ..... .... ............................................................. ................

17 Applying Solid Modelling and Animated Three-Dimensional

14 7 147 148 149 149 150 151 152

Graphics

157

Introduction................................................................................................. Preliminary experimentation...................................................................... Solid modelling......................... ...................... ............. .... ............................ Animating solid models of St.Veit-Klinglberg data.............. ..................... First video sequence................................................................................... Second video sequence........................................ ........ ...............................

15 7 158 158 159 159 160

Paul Reilly

17 .1 17.2 17.3 17.4 17.5 17.6

133 133 134 13 7 143 145

Stephen Shennan

iii

CONTENTS

17.7 Recording pictures onto videotape. ............... ............................................ 161 17 .8 Outlook... ................... .................. ........... ....... ............................................... 161 17 .9 Conclusions ............ ...... ...... ........................ .... .. ........ ...... .. .. ........ .... ........ ..... 162

18

Deconvolution

of 6.T profile curves

Olga Karousova

18.1 18.2 18.3 18.4 18.5 18.6

19

167

Milos Karous

Introduction ................................................. .................... ...... ...... ................ The construction of the filter ........................... ................................... ....... Deconvolution of 6.T profile curves ........... .................................. ............. Deconvolution filters for a plate and a cylinder. ...................... ................ Deconvolution curves ................... ....................... ...... ................. ....... ......... Conclusion... ......................................... ..... ............. ................... ....... ...........

Met hod s f o r Pro ces s ing Digit al Ge op hysic al Data in Archae o log y V. Hasek R. Vencalak 19 .1 Introduction ......................... ................. ............ ............ ................. .............. 19.2 The main trends in geophysica l data processing. ............... ..................... 19.3 Primary processing of the field data.. ....................... ...................... ........... 19.4 Qualitative interpretation ........................ ...................................... ........... .. 19.5 Conclusion ..................................................... ................. .......... ...................

Ill

STATISTICAL APPLICATIONS edited

20

Fifteen years of contributions

by C. Orton

167 167 169 170 172 178

17 9 179 179 180 180 191

193

of the French school

of data analysis

193

F. Djin dji a n

20.1 20 .2 20.3 20.4

Int r odu ct ion ....................................... ............................................... ........... T he Fren ch school of data analysis............. ................................. .............. The ap plications of data analysis techniques in archaeology ................. T he i n fl uence of data ana lysis on the evolution of archaeo l ogical met h ods ............ .. .. ........ .. ...... .... .... .... .. .. .. .. .... .. .... ...... .. .... .. ...... .. .. .. .... ........ ... 20 .5 Ot her appli cati ons...... ................... .......... ........................... .......... ............... 20 .6 Data analys is as a cognitive process ............................ ..... ............ ............ 20 . 7 Conc lusio n ................................. .......... ................................... ......... ............

21

Seriation

and Multivariate

Statistics

193 193 194 196 198 198 199

205

Torste n Madse n

22

Seriation

by similarity

R. R. Laxton

22. 1 22.2 22 .3 22.4 22.5 22.6 22. 7

23

cy

215

Int r odu ction .................... ................. ........ ............................................. ....... Some re sult s and exam pl es ...................... ......................... ................... ...... Method of seriation ................. .......................... ................................. ......... Example.. ........... .................................................. ............... .................. ........ Seria ti ng actu al data................... ................... .................... .......................... Exam p le (One room in the mound at Awatovi). ................. .............. .......... Conclusions ......................... ............... ..................... ......... ...... .............. .......

Principal Components M. ]. Baxter

23.1 23.2 23.3 23 .4

and c onsisten

]. Restorick

Analysis

of compositional

data in archaeology

215 216 217 219 220 222 224

227

M. P Heyworth

Introduction ................................ ........................................... ...................... The composition of gla ss............................ .............................. ................. . Analytical strategy ......................................... ............................. ................. Methodology... ........................................................... .............. ........... .........

iv

227 228 228 229

CONTENTS

23. 5 An application.............................................................................................. 23 .6 Archaeological considerations ................................................................... 23. 7 Conclusions................................................................................................. 24

The chemical

composition

M. N. Leese

M.]. Hughes

231 23 7 238

of tiles from Bordesley

241

]. Stopford

24 .1 Introduction..... ............. .......... .......... .... .............. ..................... .................... 24.2 Data analysis ......... ........ ....... ........................... .................. ............... ............ 24.3 Conclusions........................................................................................ ......... 2 5 Geodetic

and cartographic

problems

in archaeological

241 242 247

data bases

2 51

Irwin Scol/ar

25.l 25.2 25.3 2 5 .4 25.5 26

Maps, archaeological databases and their uses ......................... ........ ..... Mapping methods ............ ................. ....................... ............. .......... ...... ... ... Transformations from geographic to grid reference and inversely ....... Searching databases containing grid references ........... .... .... ...... .. ...... ..... A map independent storage convention for site and find recording for data exchange .............................................. ......... ........ ..... ........ ...........

Error structures Clive Orton

26.1 26. 2 26.3 26.4 26.5

of ceramic

assemblages

267 2 75

Paul Tyers

Background ................................................................. ................................. Statistical theory ........ .... .......... .... .. .. ...... .. .. .. ....... ..... .... .... ........ .. .. .... .... .... ... Data .............. ... .... ............. ........ ...................................... .... ... ............ ........... Results ........................................ ........................... .............. .... ......... .... ........ Discussion ................ .......................... ..... ................ ...... ...............................

IV RECORDING SYSTEMS edited by J. D. Richards 2 7 The development

251 252 257 264

of an integrated

275 2 76 281 282 282 287

archaeological

software

system

287

]. Huggett

27.1 27.2 2 7. 3 27.4 2 7. 5 27.6 27.7 28

Introduction ...................... .......................... ... ..................... ................... ...... Integrated software ...... ............ .................... ............... ............................... Integrated software in archaeology ........ ........... .... .... .... ...... .... .. .. ..... . .... ... Constructing an integrated system from commercial software ............ . An example system - the Dean sway Archaeology Project ...... .............. .. . Conclusions ........................... ......... .... ............................................. ..... ...... The Future ................................... ...................................... ..........................

The COMPASS System Andrew

295

Weiss

28.1 Introduction .............................. ............................ ...................................... 28.2 System Configuration ......... ........................................................................ 28.3 Applications ............................................................................................... 28.4 Graphic overlays ......................................................................................... 28.5 Stratigraphic recording at a Near Eastern tell .......................................... 29

287 287 288 289 290 291 292

Archaeological

computing

in South-Western

Europe

295 296 . 303 309 314 319

Daniel Arroyo -Bishop and Maria Teresa Lantada Zarzosa

29.1 29.2 29.3 29.4

Introduction .................................. ............................................................... France .......... ..................................................... ...... ..................................... Spain ........................................................................................... .................. Portugal ......... ... ..... ......... ....... ............... .... ...... .............................................

V

319 320 321 323

CONTENTS

29.5 Principality of Andorra .............................................................................. 29.6 Addresses..................................................................................................... V TRAINING & EDUCATION edited 30

Hypercard

as a teaching

Clive Orton

by S. P. Q. Rahtz

323 323 327

tool

32 7

Roger Grace

30.1 Introduction-what is Hypercard? .................... ............ .... ...... ...... .. .... ....... 3 2 7 30.2 Uses at the Institute of Archaeology.......................................................... 327 30.3 Discussion.................................................................................................... 334 VI EXPERT SYSTEMS AND ARTEFACT CLASSIFICATION edited 31

Beyond classification Jitu Patel

31.1 31.2 31.3 31.4 31. 5 32

by G. Lock 339 3 39

Arthur Stutt

Introduction ................................................................................................ . 339 Model of archaeological interpretation ..................................................... 339 KIVA-the archaeological interpreter......................................................... 342 Looking ahead.............................................................................................. 343 Conclusion................................................................................................... 346

Shape Information

in an artefact

Database

349

K.]. Goodson

32.1 32.2 32.3 32.4 32.5 33

Introduction................................................................................................. Shape acquisition............................................. ........................................... Shape representation .................................................................................. Shape storage....................................................... ........................................ Conclusion ............................................ .......................................................

The analysis

of ancient

Gao Liming

33.1 33.2 33.3 33.4 33.5 34

Chinese

Luo Hongjie

pottery

and porcelain

shapes

363

John Wilcock

Introduction................................................................................................. Pottery and porcelain forms to be analysed............................................. Methodology......................................................................................... ....... Results .......................................................................................................... Conclusions .................................................................................................

Rule-based

349 349 355 356 357

Dating of Artefacts

363 364 364 370 373 3 75

Kazumasa Ozawa

34.1 34.2 34.3 34.4

Introduction................................................................................................. Modelling..................................................................................................... System implementation.................................. ............................................ Conclusion...................................................................................................

vi

3 75 3 75 3 79 384

1 Preface Sebastian Rahtz* Julian Richardst Over 150 archaeologists,

scientists attended the 1989 held at the University of York from 21st - 23rd March . For the first time in the history of CM the organisation of the conference was a collaborative venture involving two bodies : the Department of Archaeology at York University, and York Archaeological Trust. The main theme of CAA89 reflected this collaboration, with a major emphasis upon the application of computers and statistics by field archaeologists. The conference was also notable for the large number of overseas delegates who attended, including visitors from France, Germany, Holland, Italy, Japan and the United States. Many individuals helped to make the Conference a success, but our co -organisers from York Archaeological Trust, Jef Maytom and Karen Torevell, deserve a special mention. Throughout the year the CAA steering committee, under the chairmanship of Clive Orton, kept a careful eye on progress. Once more Sebastian Rahtz has played a leading role as production editor in ensuring the rapid publication of a volume of papers which reflects the conference. Our policy, this year, has not been to produce a complete record of the conference. Many speakers have preferred not to offer their papers for publication, whilst other contributions are being published elsewhere. For the present volume, we have chosen to concentrate on five main themes, each of which was well represented at the conference: cultural resource management, excavation recording, surface and solid modelling and image enhancement of archaeological data, statistical applica tions, and expert systems and artefact classification. Members of the CM Steering Committee, namely Ben Booth, Gary Lock, Clive Orton and Clive Ruggles have kindly acted as theme editors, and to them we owe considerable thanks. We are grateful to the contributors for putting up with some harsh deadlines and conditions of submission, and to Kiran Mistry at the University of Southampton for undertaking the retyping of papers which did not arrive in machine -readable form. Most papers were sent on floppy disc or via electronic mail, and most of the text was used direct, although mathematical formulae and tables have often been recast entirely. Authors were encouraged to submit machine -readable figures and a gratifying number arrived in this form, speeding up production. Other figures were pasted in manually, but in a number of papers the originals were digitally scanned, and the resulting bitmaps automatically included in the page makeup process. Computer Applications

*

mathematicians,

and Quantitative

and computer

Methods in Archaeology conference,

Department of Electronics and Computer Science University Southampton S09 SNH Department of Archaeology University of York Micklegate House Mickle gate York

vii

2 Investment appraisal for information technology-the experience of the English Heritage record of scheduled monuments Nigel Clubb* 2 .1

Introduction

In both the public and the private sectors, there is an increasing emphasis on the need to identify the costs and benefits attributable to alternative methods of meeting objectives. This process can be used to assess and justify computing proposals, not merely as a test of financial viability, but also as a basis for decision -making and the monitoring of a project. The decision -making must be related to the aims of an organisation and how its computing resource should be exploited to achieve those aims. It must also be related to information technology strategy. In the civil service and elsewhere, simple techniques of investment appraisal (IA) have been evolved to approve and monitor projects and these can be applied to computing proposals of any scale. They include a total assessment of cost over the life of a project, (and it sometimes surprising how these can mount up even for quite modest systems), and ways of assessing the quantifiable and unquantifiable benefits of a proposal. Hardware and software are not usually the major cost compared with staff time on input and editing, consultancy, maintenance, accommodation, training and consumables over a period of years. Some of the questions which may be asked to test the appraisal are what risk is there that the system will fail to provide the expected benefits, or, are the benefits sensitive to increased costs or the loss of key members of staff? What level of risk or speculation is acceptable and what is the level of probability? Is all the expediture 'upfront' with the benefits arising in five years' time and what are the implications for cash flow? In this paper, I will define briefly the process of IA as defined by the Treasury, refer to the wider context of procurement decisions and summarise IA as applied by the decision to provide a new computer -based record of scheduled monuments at English Heritage.

2.2

Investment appraisal as defined by HM Treasury

IA is defined by the Treasury guidelines (HM Treasury 1984) as a systematic approach to expenditure decisions which entails deciding clearly on the objectives, the various ways of meeting them and working out and presenting the costs and benefits of each option. It seeks to question and challenge what is being done and sets the tone for rational thought about the use of resources. It does not eliminate risk or speculation, indeed, it may show that a greater risk than originally proposed is worthwhile or that greater expenditure will produce greater returns. The sequence often followed is; * English Heritage Fortress House 23 Savile Row London WlX 2HE

1

NIGEL CLUBB

1. To define the objectives of a project, 2. To consider the options for meeting those objectives, of which one may be to do nothing and examine the full consequences and another to attempt the maximum possible, 3. To identify the costs, benefits, timing and uncertainties

of each option,

4. To discount those costs and benefits which can be quantified in money terms and present a net present value (NPV) for each option, 5. To state and analyse the qualitative benefits, if genuinely unquantifiable, (at the very least, unquantifiable benefits should be analysed in terms of standards of service), 6. To weigh up uncertain factors and assess any other relevant issues, 7. To present the results for decision -making.

2 .3 The context of investment appraisal In this paper I shall be arguing the IA is not just a bureaucratic device but that, used properly, it can assist in determining solutions and monitoring the extent to which projects satisfy their original expectations, particularly in terms of benefits. In the case of the English Heritage computer-based mapping system, (Clubb 1988), an IA demonstrated that the project needed to be expanded to include listed buildings as well as scheduled monuments in order to realise maximum benefits. IA alone cannot justify all projects since there are also human, intuitive and cultural dimensions. IA is only part of a wider process and not always the most important part.

2.4

Investment appraisal and 'willingness to pay'

In the past, it has been much easier to regard 'heritage' benefits in terms of service alone, rather than the financial value of benefits. However, service has an associated cost which must be met, even in the public sector. The use of information technology often seems inevitable in an information age. The decision to embark upon a computer application is usually at least partly based on a feeling that a computer solution is correct in terms of the quantity of data to be handled, or the complexities of information handling, or the calculations required or the numbers of users involved. However, all organisations and activities, even universities and pure research, may be subject to externally imposed measures of performance, often of a crude nature. 'Gut' reactions and professional judgement on the one hand and performance indicators on the other may seem miles apart, but they tend to come together at the level of 'willingness to pay' . To take one example outside the world of archaeology, how much should British Rail spend on safety, bearing in mind that the more spent on safety the less is available for station modernisation and new trains etc? It has been argued (Jones -Lee 1989) that in purely economic terms safety improvements should only be carried out as long as their value exceeds the cost. We can only estimate this if we have explicit costs and benefits. It is relatively easy to estimate the costs of specific safety improvements. I understand that the Department of Transport have traditionally set a value on human life in terms of output through working career . However, most of us value safety because of an aversion to death or injury rather than a desire to protect future

2

2.

INVESTMENT APPRAISAL FOR INFORMATION TECHNOLOGY

earnings. Ultimately, therefore, the value of our own lives to ourselves depends on how much we are willing, or able, to pay to avoid a transport injury. Another example a little closer to home concerns the joint proposal of English Heritage and the Royal Commission on the Historic Buildings of England for a national computerised data -base of listed historic buildings. In 1986, the Environment Select Committee of the House of Commons (House Of Commons 1987) expressed surprise that details of the 400,000 or so listed historic buildings in England were not held on a computer data-base. In the following year, a joint English Heritage/RCHME report (unpublished) to the Department of the Environment (DoE) dealt with the costs and benefits of such a data -base. The DoE has consistently taken the view that the costs of such a system should be carried by those who benefit from it. Clearly, DoE themselves, English Heritage and RCHME could be considered to come into this category, as, perhaps, could local authorities. However, if the net is widened, there may be a number of commercial interests who might gain benefit from s ch a data base, including publishers of magazines which exploit interests in pe riod houses and 'country' lifestyles and their advertisers who wish to target the readers of such magazines. Certain types of developer and estate agent might also be interested. Current (1989) feasibility studies into this project are likely to take account of (and possibly carry out market research into) the possible value to such organisations and the likelihood of contributions from them. Both the examples quoted here show that the value of activities may be some distance away from those providing the service. This has implications in terms of 'willingness to pay'.

2.5

The English Heritage record of scheduled monuments

The existing data -base for records of scheduled monuments (RSM)has been discussed (Booth 1988). As early as 1986 it was recognised that a complete replacement would be required in order to service the scheduling process and the number of additional scheduled monuments anticipated in the course of the Monuments Protection Pro gramme (MPP). Indeed, the existing computer system is approaching the end of its useful life. The main outputs required from the new system are as follows; 1. Documentation

arising from the scheduling process, ie the generation and cir culation of scheduling proposals, preliminary notification letters, recommen dations to the English Heritage Ancient Monuments Advisory Committee and Commissioners, preparation of recommendations and associated documenta tion and the generation of the published lists of monuments .

2. Generation of reports to English Heritage Inspectors on the receipt of new management information following the visits of Field Monument Wardens to monuments. 3. Generation of statistics and other data on site management in order that overall resource requirements and priorities for management action subsequent to scheduling can be identified. 4. Reports from the data -base required in the course of case -work, eg distributions of particular monument types, monuments under threat, monuments under management agreements. 5. Output, mainly in magnetic form for data -transfer to RCHME and county -based Sites and Monuments Records.

3

NIGEL CLUBB

2.6

Net Present Value

The analysis of costs and benefits is concerned to see if projects are worthwhile from a financial point of view. Almost all projects produce benefits later than costs. However, the value of money spent today tends to fall compared with money spent later. There may be balances to be struck, eg between increased investment now and increased running costs at a later date. In order to estimate the value of a proposed investment and its costs over a protracted time -scale, compared with other options, Discounted Cash Flow or Net Present Value (NPV) can be used to bring all the data back to one common point of time, ie to work out the present value. Discounted costs can then be subtracted from discounted benefits. The familiar basic discount formula (OF) is expressed to give a factor by which to multiply the original figure by, hence; DF

where r years

2. 7

=

rate of discount

Quantifiable

expressed

=

LO (1.0 + r)

t

as a decimal (eg 5% = 0.05) and l

=

number of

costs and benefits

The project costs are set out in Table 2 .1, ie non -recurrent costs of £200,900, recur rent costs building up to £57,616 pa and a total project cost of £605,380 over project life. The discounted cash flow(DCF) is shown in Table 2.2 which gives a OCF of £537,142, after the application of the OF based on Treasury discount tables. The quantifiable benefits arise from two main sources. Firstly, the administrative processes associated with MPP require increased administrative staff. On a worst 'scenario' basis, it is estimated that a 10-15% increase in productivity is possible. This amounts to the equivalent of four staff who it will not be necessary to deploy for this purpose. These benefits are shown under the heading 'scheduling' in Table 2.2. Secondly, the supply of information from the system in the course of case -work is estimated to lead to a 5% increase in productivity, (shown as AMO in Table 2.2). These benefits are quite considerable over the life of the project, ie valued at £757,184, discounted to £612,673. Consequently, the quantified costs and benefits taken together produce a NPV of £75,531. This is not a vast return on the original investment, but is, at least, a positive figure.

2.8

Alternative solutions

Three main alternatives were investigated, ie a fully manual system, a micro -based system, (largely the 'status -quo'), and a system of networked micros controlled by a central file -server. The fully manual system involved dispensing with the exisiting computer -based record and the various mica -computers already in use for office systems. The additional staff required to provide a basic service over a seven -year period were estimated to cost about £3m and the option was not considered viable. A more realistic option was a micro -assisted project. This was very largely the 'status-quo'. which provides for word -processing, basic data -bases and simple mon itoring of progress. This option failed to make any additional improvements in staff productivity and resulted in a negative NPV of - £384,057, in effect most of the project costs since the productivity arising from the 'status quo' had already been taken into account in the staffing levels provided for MPP. 4

2.

INVESTMENT

APPRAISAL FOR INFORMATION

TECHNOLOGY

£

I. Non -recurrent costs

l 3,400

Terminal x 12 (local) Terminal support equipment Printers x 4 (local) Auxiliary storage (600 mbytes) Processor Tape Drive Console printer Cabling etc Operating software Application software Software development

5,600 1,800 11,700 59,300 19,200 1,400 500 23,000 32,000 33,000 200,900

Total N/R

YI 25,112

Y2- 7 25,112

Nil 8,168 5,000

5,000 24,504 3,000

5,000 25,504 1,000

13,168

57,616

55,616

2. Recurrent costs

YO

Maintenance Consumables Additional staff Training

Nil

Total R

3. Total project cost (Year 0-7) = £605,380

Table 2.1: Project cost Diacounted

ca•h flow 88/89 0 £187 ,900 {13 , 168

Noncurrent Rr,currt,nt

£210 ,068

Total costs Cumulative

cost

DF DCf' Cumulative

leneftta

DCF

£201 ,068

89/90 I {13 ,000 {57 ,616 {70 ,616 {271,684

90/91

91/92

92/93

2 £0 { 55, 616

3 £0 {55 ,616

£0 {55 ,616

{ 55, 616 £327 ,300

1.000 £201 ,068

0 .952 [67,226

[201 ,068

[268 ,294

0 .907 £50 ,444 £318,738

{55 ,616

4

93/94

94/95

95/96

5 £0 £5 5,616

6 £0 { 55, 616

7 £0 {55 ,616

{200 ,900 {404,480

{55 ,616

{5 5,616

{ 55, 616

{ 55, 616

{605,380

£438, 532

{494 , 148

{ 549 ,764

£605 ,380

{60 5,3 80

0 .864 [48 ,052

0 .823 [45,772

0 .783 [43 ,4 5 7

0 .746 [41 ,490

0 .711 £39 ,5 43

£ 53 7, 142

£366 ,7 90

£412 ,562

£456 , 110

£497 , 599

£537 , 142

£5 3 7, 142

£382,916

and Nl'V 0

I

2

91 /92 3

£0 £0

£36 , 118 £0

£72 .236 £26 , 150

£72 ,236 [52,300

Tota l Benefits

£0

£36,118

£98 ,3 86

£124 ,536

Cumulative

£0

£36,118

[134 ,504

[259 ,040

0 .000

0 .952

0 .907

0 .864

0 .823

£89 ,236

{107 ,599

£ I 02 .493

{97 ,5 12

£9 2,90 4

{231 ,220

£333 ,713

{431 ,22 4

£52 4 , 128

£56,721

£53 ,964

£51 ,414

88/89

Scheduling AMD

Benefits

DF

89/90

90/91

Discounted

Br,nefits

co

£34 ,384

Cumulative Benefits

Discounted

£0

£34 ,384

Discounted Benefits

net

(£201 ,068)

(£32 ,842)

MPV -

TOTAL

£123,620

£38 ,7 92

[ 59 ,5 47

92/93

93/94 4

[72 ,236 £52,300

5

5

7

TOTAL

£469,534 £287 ,650

{124 ,53 6

£124 ,53 6

£757,184

{632 ,648

[757 , 184

{7 57, 184

0 746

0 .711

£72 ,236 [52 ,300

{124 , 536

£124 ,536

£383 , 576

£508 , 112 0 .783

cash flow

95/96

[72 ,236 £52 ,300

£72 .236 £52,300

[75,531

Table 2.2: Discounted

94/9 5 6

{88 ,54 5 £612,673

[49 ,002

{612 ,673 [612 ,673

[7 5,5 31

NIGEL CLUBB

The option of a network of micro -computers linked by a central file server merited more detailed consideration. This would provide office automation facilities and the use of elementary data -bases to handle indexing and searching. Local micro computers would be connected via a local area network to a high capacity disk storage. The option produces some of the benefits of the preferred option, ie those resulting from the scheduling process, but not those associated with case work. The costs of this option over 7 years were quite high, (£497,950, discounted to £438,204). The scheduling benefits were £469,950, discounted to £383,573, producing a negative NPV of -£54,631. There was concern that the option would provide a limited service without any significant interactive use by those needing to consult the records or to drive the administrative processes associated with scheduling and case -work.

2.9

Unquantifiable

The unquantifiable

benefits

benefits of the project are summarised

as follows;

1. A complete, accurate and updatable base of data is essential to ensure the effec tive application of scheduling legislation and the management of monuments with statutory protection. 2. It will provide the opportunity for action to be taken in the light of knowledge of the state of scheduled monuments and their management, including statis tics. It should lead to more effective policies for monument management and conservation in the light of priorities for action. 3. It should lead to the better provision of information to Government ministers, the DoE and local authorities on the stock of scheduled monuments and their characteristics. 4. Automatic transfers of data between local authorities, RCHME and English Her itage will encourage more integrated conservation computing and record -keeping . 5. There will be opportunities

to monitor staff output and performance.

6. There will be opportunities to develop confidence and expertise among staff and a contribution towards staff morale and to enhanced effectiveness and reduction of routine clerical work. Some of the unquantifiable benefits could be subjected to cost -effectiveness ap praisal, but only in terms of an investment appraisal on the value of conservation itself. In general, unquantifiable benefits should be subjected to some form of cost effectiveness assessment, at least in terms of the standard of service provided, expressed in financial value, wherever possible.

2.10

Uncertainties

The next stage in the IA process was to consider the uncertainties on the basis that activities rarely turn out as expected. The main potential uncertainties for the RSM were whether the savings in staff productivity would be realised and whether the software development costs had been accurately estimated. In addition, staff costs may. or may not, keep pace with inflation. The less staff are remunerated over time, the less the financial savings arising from productivity benefits will be. Uncertainty may move in two directions since the benefits may be more or less than expected and 6

2.

INVESTMENT APPRAISAL FOR INFORMATION TECHNOLOGY

many computer projects have resulted in advantages not considered at the time of justification. There are two main approaches to uncertainty, ie risk analysis and sensitivity. In risk analysis, the aim is to quantify the potential differences between the esti mated outcome against a range of expectations from the optimistic to the pessimistic. The object is to force project leaders to think about potential problems in advance and to identify the critical elements determining the success or failure of the project and the probability of those risks actually occurring. Also, the optimistic view may identify benefits not thought of, perhaps a long way from the original project boundaries, and these may help to establish who is receiving value from the system . Sensitivity testing around the central assumptions of a project is intended to quantify the effects of uncertainties on the outcome, firstly by identifying the key elements and then varying them to calculate a series of new overall outcomes. An example of the approach is that the rate of inflation on wages is currently higher than increases in the costs of computing equipment. It may be necessary to apply differential inflation rates to different elements of the options . This tends to favour machines rather than employees, depending on assumptions about exchange rates affecting the prices of imports . In calculating the NPV for the English Heritage system we had already built in an element of risk analysis since the costs and benefits had been calculated on the basis of the most pessimistic 'scenario' which would result in a positive NPV. The main areas of uncertainty evaluated in detail included the following; 1. The solution pro .posed did not involve 'leading edge' technology . There was no reason to believe that the requirement was not achievable through the solu tion proposed or that the computing environment would not be supported by suppliers in the foreseeable future. 2. Most of the proposed users of the system had at least some keyboard experience and no significant user resistance was expected . 3. The requirement sis.

had evolved during a period of 18 months thought and analy -

4. The main area of uncertainty related to the costs of software development. It was decided that any additional expenditure here must be subject to a separate IA for which the Project Team would take responsibility in the first instance .

2.11

Conclusions

The project was approved by the Treasury in November 1988 . The process of securing approval had taken about 12 months and this paper does not pretend to deal with all the issues and analysis carried out. The IA was a requirement of the Treasury and all aspects of the system were subjected to considerable scrutiny. In many respects, the IA did not produce any new evidence to be taken into acccount in the procurement decision, particularly since a viable alternative to the proposed solution did not emerge. Also services such as improved management information cannot be valued in purely financial terms. In addition, in a world of uncertainty, intuitive judgement by those with experience is highly relevant. However, English Heritage will be able to undertake the development of the system with a full awareness of the costs and their implications, together with the potential risks and pitfalls and can take management action to avert or mitigate them. The IA demonstrated that the financial benefits were relatively sensitive to varia tions in the levels of productivity actually achieved. The extent to which English Heritage was prepared to finance any shortfall (ie a value which is less than the 7

NIGEL CLUBB

investment) depended on the unquantifiable benefits and the strong view of those most directly concerned as users and curators of the data that the system was essential and that the relatively small financial benefits had been calculated on a pessimistic interpretation of likely events. The test of the accuracy of the IA will emerge in the course of the project im plementation. However, it has already provided a firm base for the monitoring of the project. This will include the direct costs and benefits and the major state of the world asumptions against which the original proposal was framed. There will be a need for post -implementation reviews and retrospective appraisals in order to ensure that best value is secured for the cost and that the project is questioned and challenged and continues to receive rational thought and planning . It will also ensure that lessons are drawn from any mistakes made. It is a requirement of human psychology that people at work need to feel justified and valued, just as in other areas of life. Disciplines such as archaeology will be increasingly challenged unless an attempt is made to place value on the services provided, particularly since they consume resources provided by someone. The costs of services and their relationship with potential paymasters are coming into sharper focus as certain types of archaeological s ervice are privatised, (or partially so). Methods of evaluating such service outputs are being refined and improved constantly and the profession should be prepared to embrace them to ensure that funds for continuing activity are forthcoming. ·

Bibliography BOOTH, B. 1988. "The SAM record - past, present 3 79- 388. CLUBB,N. 1988. "Computer 408.

and future". In Rahtz 1988, pp.

mapping and the SAM record". In Rahtz 1988, pp. 399-

HM TREASURY1984. "Investment Appraisal for Government Departments", London .

in the Public Sector, A Technical Guide

HOUSEOF COMMONS1987. "First Report from the Environment buildings and Ancient Monuments", London.

Committee - Historic

JONES-LEE, M. 1989. The Economics of Safety and Physical Risk. RAHTZ, S. P. Q., (ed.) 1988. Computer and Quantitative Methods in Archaeology 1988, International Series 446, Oxford. British Archaeological Reports.

8

3 Computerised county sites and monuments records in England-an overview of their structure, development and progress Amanda Chadburn * 3 .1

Introduction

During 1987 and 1988, the Records Office of English Heritage undertook a survey of the 46 county -based Sites and Monuments Records (SMRs) in England as part of the preparation for the Monuments Protection Programme. Visits were carried out by the author, Ben Booth, Nigel Clubb, and Bill Startin. Whilst details of the survey are confidential, some overall trends have emerged, and these are discussed in general terms. It should be noted that these visits took place between September 198 7 and the summer of 1988, and that most SMRs will have progressed since then. It is hoped, however, that as an overview it will be useful for showing trends and developments within SMRs.

3.2

The history of SMRs

SMRs are generally the most comprehensive record of archaeological sites within a county. They usually consist of records of individual archaeological sites or finds, which contain standard pieces of locational, administrative, and descriptive information, including details of the site -type and period of the archaeological item involved. Some SMRs also include a section for management recommendations on archaeological sites. Each text -based record has a unique reference, sometimes called the Primary Reference Number (PRN). Text -based records are accompanied by a set of location maps which show the positions of sites or findspots, referenced to the text records by the PRN. SMRs in England are often based in County Council Planning Departments where they act as an aid to the planning process. The first was started in Oxfordshire in the late 1960s, and many followed in the early 19 70s, most being originally set up as paper -based record systems (often card -based systems). SMR cover in Wales and Scotland is structured slightly differently : Wales has four archaeological trusts, and five SMRs, whereas the cover in Scotland is much more patchy. This paper will only discuss the SMRs in England. SMRs have developed successfully and very rapidly over the last few years. In the early 1970s only a few counties had any cover; by the late 1970s, most counties had an SMR, although many were in the early stages of data compilation. Today, all the counties in England have SMRs, which are either computerised, or in the process of computerisation. This is a significant achievement for the archaeological profession, as it allows informed decisions to be made about our cultural heritage, and provides a tool for a range of activities such as education, research, and planning. * English Heritage Fortress House 23 Savile Row London WlX 2HE

9

AMANDA CHADBURN

English Heritage, and the DoE before it, have been funding SMRs in advance of the Monuments Protection Programme. This programme is reviewing all known archaeological sites in England, and aims to greatly increase the number of scheduled or statutorily protected archaeological sites. In order to do this, full and retrievable records of all known archaeology in England, which can provide the data for such a review, are needed. The SMRs were chosen as the records most able to provide this data. English Heritage has been grant-aiding SMRs for a number of reasons, but partly so that a relatively common standard of content and retrievability is available for the whole country. Inevitably, there is some variation in the breadth of cover, state of compilation, and computing capability of SMRs, and these variations are discussed below.

3.3

The scope and content of SMRs

Variations in the width or scope of an SMR do occur, although it has been rec ommended that ideally there should be no policy decision to exclude records on account of their date or their type' (Historic Buildings and Monuments Commission for England 1985). Most organisations allow the categories of evidence shown in Table 3.1 in the SMR, although the extent to which these have been systematically •trawled' varies. Site -types and finds of any date range are allowed in most records, although a few SMRs have cut-off dates of the late medieval period, or 1700 A.O. or 1900 A.O. Most SMRs cover their entire area or county without any geographical exclusions, although there are notable absences which include some of England's most important historic towns. Some of these urban areas are covered by separate local SMRs, but in others there dces not appear to be any SMR cover at all. 1

3.3.1

The range of information

in SMRs (the depth of the SMR)

There is a basic core of data which most organisations compile for every SMR record, effectively a •record content standard', comprising the fields of information shown in Table 3.2.

3.3.2

Rarely used fields of information

Only 6 organisations stated that all fields in their SMR records are regularly used and completed for all sites. The majority of computerised SMRs contain fields which are rarely or never used, over 13 fields per record in some cases. These usually include survival and condition, (at least 18 SMRs rarely contain these data) and owner and occupier. Other fields which are often left blank include area status, site status, management information, soils, geology, assessment of importance, and land classification or use. Such data is often omitted because sites are rarely visited, and data capture cannot take place without a visit. It is also difficult to keep such information up to date (a requirement of the Data Protecetion Act) without a regular programme of visits, and this, unfortunately, is beyond the resources of most SMRs.

3.4

Data structure

There are major differences in the conceptual structure of SMRs, relating to the way an individual SMR record is defined, and what it actually consists of. There are three main approaches taken. An individual record can relate to : 1. a single archaeological

item;

10

3.

COMPUTERISED COUNTY SITES AND MONUMENTS

RECORDS IN ENGLAND

Listed buildings Non-listed buildings Finds pots Placenames (including fieldnames) Museum or private collections of artefacts Archaeological sites, both upstanding and buri ed remains . AP Evidence

Table 3.1: Categories

of evidence allowed in most SMRs

County Local authority Parish NGR Name Description Site -type Form Period - general Period - specific Survival Condition Land class on site Land class around site Site status Area status Owner Occupier Management Assessment of importance Archaeological History

Table 3.2: Data categories

within a typical, individual

11

SMR Record

AMANDA

CHADBURN

2. a single land parcel containing 3. a single item(s).

piece of recorded

an archaeological

or received

item(s);

information

about

an archaeological

Individual records in most SMRs are (primarily) those of archaeological items. However, sometimes a combination of approaches is used e.g. many SMRs do hold 'land parcel' records and 'archaeological item' records, to avoid the problems inherent in recording such sites as cropmarks, extensive or dense remains, and urban areas by archaeological integrity or item. Only one SMR uses the 'land parcel approach' exclusively, i.e. all archaeological sites falling within a specified area will be documented in a single record. There are further SMRs who have stated that they would like to make more use of the 'land parcel approach' in future. No SMR contains records of 'received information' alone, but sometimes this ap proach is used in conjunction with 'land parcel' or 'archaeological item' records. One SMR uniquely uses a hierarchical system where primary records relate to an archaeological item, and component records relate to received information/event, e.g. a primary record for a Deserted Medieval Village, will have component records of every visit, survey, excavation and so on, for that site. 27 SMRs use component records, although these vary greatly in structure and content.

3.4.1

Controlled vocabulary;

its use and origin

The subject of vocabulary control has been discussed by the author elsewhere in depth (Chadburn 1988). Most SMRs use some form of vocabulary control, at least for site -type and period, although many others also use controlled vocabulary for land classification, local authorities, site status, area status, form, etc. Some SMRs have either developed their site -type vocabulary in -house, or use wordlists based on vocabulary in use at English Heritage, the Council for British Archaeology, or the Royal Commsission on the Historical Monuments of England (or use a mixture of all of them). Occasionally, SMRs do not use controlled vocabulary for site -type, simply entering any appropriate terms into the record, and then producing wordlists of site -type terms used. Searches are carried out using selected terms from the resulting wordlist, which has the disadvantage that a search for a particular object or site -type may have to be carried out a number of times.

3.4.2

Manuals for staff and the documentation

of the SMR

Generally speaking, most of the SMRs visited have some sort of accompanying manual, although the majority of manuals relate only to the compilation of the computerised record, rather than documenting any strategic framework. Most have been written specifically for the compilers rather than for users of the SMR. There are some exceptions to this, and several SMRs have wide -ranging and carefully doc umented manuals. However, most manuals are limited in their scope and therefore, usefulness.

3.5 3.5.1

Back-up material Associated map systems

Most SMRs hold two basic sets of map cover for the county: the base maps, and constraint maps. Sites and findspots are plotted onto base maps (either directly onto the maps or on film overlays) and are referenced by the PRN to the text records.

12

3.

COMPUTERISED

COUNTY

SITES AND MONUMENTS

RECORDS IN ENGLAND

Constraint maps show archaeologically sensitive areas, and a set of these maps is often held District/Borough Planning Departments to aid their decision -making processes. Base maps are usually at 1: 10,000 or 6' scale, although sometimes additional maps at a large scale (at 1:2500, 1:1250 or 25') are held for urban areas, or areas which have a high density of sites. Constraint maps are usually held at the same scale as the base maps. Occasionally, a much fuller set of maps is held with an SMR. These include maps of a variety of scales and dates e.g. the O.S series first edition maps, copies of tithe maps, copies of sale maps and numerous overlays to the base maps, which might show features such as medieval field systems or field names. However, this level of associated mapped information is rare in SMRs.

3.5.2

Associated record systems and back-up information

Most SMRs hold some documentation in addition to their card systems, or comput erised files. Not unnaturally, long established SMRs, such as Bedfordshire, Oxford shire and Buckinghamshire, are usually the fullest and contain the most associated information. However, many SMRs hold only a few examples of the following items : • Journals • Offprints • OS Cards • AP Collections • Unpublished

research

• Rescue and/or

assessment

surveys

Some of the long established SMRs have a thorough coverage of the complete county for most or all of the above items, although the newer SMRs tend to hold little associated information. In addition to the above items, many SMRs also hold parish and/or site files, which contain newpaper cuttings, correspondence, sketch plans etc. relating to archaeological sites or to the archaeology of an area.

3.5.3

Relationship to listed building records

Few SMRs actually contain a complete computerised record of listed buildings for the county. Some organisations hold non -computerised records, or copies of the 'greenbacks' as associated information to the SMR. Generally, most SMRs do not contain any records of listed buildings, apart from those which were (exceptionally) recorded by the Ordnance Survey on the OS cards, or which are also Scheduled Ancient Monuments. Some organisations have a long term aim to include listed buildings in their SMRs, but have not yet started to add them sytematically. Lastly , some SMRs are held within a planning department, which holds listed building records elsewhere in the department.

3 .6 3 .6.1

Computerisation Hardware

Of the 46 SMRs in England, 3 3 are microcomputer based, and 13 are on mini computers or mainframes (Fig. 3.1). Amongst the microcomputer based systems 25 use the Superfile package (2 are yet to upgrade from EH 'Version l '), 6 use dBase III, 1 uses MOBS and 1 uses Rescue (Fig. 3.2). Of the mainframe systems 4 use the North 13

AMANDA CHADBURN

Yorkshire System, 2 use Mapper, and there are 7 other systems in use (Fig. 3.3). Most SMRs are generally satisfied with their hardware, although some complained of a lack of speed, and there are one or two machines which are not effective for the software.

3.6.2

Software

The hardware and software in use in both micro -based and mainframe based SMRs at the time of the survey are given in Table 3.3 below. Characteristics of the four main packages in are shown in Table 3.4 below. Most SMRs are generally satisfied with their software, although some limitations were reported, e.g. in Superfile, the lack of speed, the size of the record allowed, and the poor documentation from the suppliers. Some users noted that the lack of repeating fields in rJBaseIII was a problem, and the users of Mapper felt that the short records required and the lack of facilities for free text were also limiting.

3.6.3

Operating systems

With the exception of Bedford (using CP/M) all PC systems use MS-DOS. The main frame/mini systems use operating systems proprietary to their hardware. There are currently no SMR systems using Unix.

3.6.4

Networking

Only fopr SMRs are networked at present, 3 are micro -based and one is main -frame based. Those using microcomputer -based networks have experienced performance problems in the past.

3.6.5

Data input, validation, back-up and archiving

14 organisations enter their SMR data in batch mode ; 20 enter data on -line and 12 use acombination of the two. 15 SMRs do not have any automatic data validation facilities, although with the exception of two SMRs, (who do not check their input data by any method) data is manually checked after entry. The back -up and archiving of SMR data is generally satisfactory. In the case of main-frame systems, this is done centrally, but most of the micro -based SMRs are careful to have a number of back -up copies of their data, usually with at least one copy off -site.

3.6.6

Output and retrievability

Only a few computerised SMRs cannot yet retrieve anything, due to operating diffi culties or non -operational computer systems. However, there are others who have not fully recast their paper records onto computer, and who thus cannot retrieve on all known sites automatically. The SMRs which are still in the process of data compilation or computerisation, are mostly in receipt of grants from English Heritage to aid the continued recasting and computerisation of data. A few SMRs still rely heavily on Optical Co-Incidence Cards for retrieval, and they can only undertake limited searches of their records . Additionally, there is the problem that any output from the Co-incidence Cards cannot be automatically produced, and neither can it be automatically transferred to another database should this be required. Most SMRs routinely produce print -outs ordered by parish, grid reference, site -type and sometimes period .

14

3.

COMPUTERISED COUNTY SITES AND MONUMENTS

RECORDS IN ENGLAND

Minicomputers/Mainframes (28%)

Microcomputers (72%)

Figure 3.1 : Hardware

in use in SMRs, 198 7- 8

Superfile (76%}

dBasell and Ill (18%}

Figure 3.2 : Software in use in micro -based SMRs, 1987 - 8 N Yorks (31%)

Mapper (15%}

Other packages (54%)

Figure 3.3: Software in use in mainframe -based SMRs, 1987 - 8

15

AMANDA

CHADBURN

aero

users

ase

IBM PS/2 Model 60 MS DOS Avon Ferranti MS DOS Cheshire PS/2 Model 60 MS DOS Cleveland North Star Dimension MS DOS Cornwall Opus MS DOS Durham IBM PS/2 Model 60 MS DOS East Sussex IBM AT PC DOS Essex Victor 286c MS DOS G. Manchester IBM AT MS N[T Novell Gloucester IBM XT PC DOS Hampshire Hereford and Worcester Tandon PCA MS DOS [riccson MS DOS Isle of Wight IBM AT PC DOS Lanes Compaq 386 MS DOS Lincolnshire Compaq Desk pro 386 PC DOS Norfolk IBM XT PC DOS "Northum . Convergent Technology N GEN MS DOS Nottingham IBM XT PC DOS " Shropshire Olivetti M28 MS DOS Somerset Staffordshire Compaq Deskpro 386 MS DOS Apricot 386 MS DOS Suffolk ICL PWS MS DOS Tyne and Wear IBM XT PC DOS West Midlands West Sussex IBM AT PC DOS IBM XT PC DOS West Yorkshire (" using English llerita ge 'Version 1' Software)

aero

ase

ystems-

Hertfordshire Humberside Merseyside South Yorks . Warwickshire Wiltshire

ase

an

users PC DOS MS DOS MS DOS MS DOS MS DOS MS DOS

IBM PC XT 286 N[C /\PC Epson Amstrad PC l 5 12 IBM Clone Compaq 386

MOBS Rescue

CP/M MS DOS

users VM[/B VM[ VM[/B VM[

ICL 3980 ICL ICL 2900 ICL

ase systems-ot

ers

users

IBM 4341 ICL 2988 Derbyshire Devon Dorset G. London Kent Lei cs Surrey

Table 3.3: Hardware

Stairs Quickbuild, Qucry mastcr, Report Master COS, SAS Prime Information SPSS,Deke /\DARAS, N/\TUR OR/\CLL: Mapper Mapp er

IBM 303 1 Prime 9650 ICL IBM D[C VAX Sperry l l 00 Sperry 1100

and Software

VM Primos VM[ AL VMS MU MU

in use in County -based SMRs in England,

I at dBase Ill

Relational

N. Yorks (filetab with Cobol enhancements ; also has links to dBasc Ill in N. Yorks CC) Mapper

1licrarchical

Table 3.4: Characteristics

VM/CMS VML:

I licrarchical

of the software

1987-8

aria e cngt rccor and groups of fields Fixed length records, repeating fields arc kept in separate files . (But NB these fea tures arc not usually used by SMRs.) Fixed length records, with repeated infor mation being handled by linked records) . Fixed length ( 132 character) records, without repeated fields . Records may be linked

packages

1987-8

16

most frequently

used by SMRs

3.

3 .6. 7

COMPUTERISED

COUNTY

SITES AND MONUMENTS

RECORDS IN ENGLAND

Data transfer

Data transfer is a very different problem to retrieval. Organisations may be able to retrieve SMR data from a bespoke computer program fairly efficiently, but they can have severe problems in exchanging it. As a general rule, those SMRs which computerised early, seem to have the most problems in both computerised data retrieval and data transfer. In addition, the Optical Co-Incidence Cards are often not kept up to date, once a recasting exercise of a paper -based system to a computerised system has begun. This can limit the effectiveness of data searches, making them awkward and lengthy, as two systems must be searched for complete retrieval. A draft data transfer standard, jointly agreed between the Royal Commission on the Historical Monuments for England, and English Heritage, has been circulated to the Association for County Archaeological Officers and to all SMRs for comment. (Booth 1988, English Heritage / Royal Commission on Historical Monuments 1988). The purpose of this standard is to provide a facility so that archaeological data may be readily exchanged between organisations, as it will all be formatted to an agreed structure. This cuts down on a costly and unnecessary formatting between separate organisations, and the case in argued in further detail by Booth (1988. Most of the main -frame users are happy with the computer resources, although their systems could be slow at peak times. Most SMRs indicated that the level of hardware and software support was adequate to very good. Only five SMRs reported that there were problems in this area, either because the support was distanced from them, or that it was inadequate or non existent.

3.7

Compilation

to date

Well over half the SMRs have now been fully recast onto computer, although further work in basic data compilation is still required in some of them. 16 SMRs have large backlogs of paper records to computerise, 13 SMRs are transferring data from one software package to another, and two SMRs at the time that they were visited, had yet to start the computerisation process.

3 .8

Conclusions

In conclusion, there has been a significant development in the quality and quantity of SMR data over the last decade. Indeed, some areas of the country are already in the position of being able to undertake evaluation exercises of particular monument classes for the Monuments Protection Programme. The fact that SMRs are now in a position to offer such services, means that taken together they undoubtably provide the most comprehensive archaeological coverage for the whole country. Despite this, they are mainly used exclusively for planning purposes, and thus represent one of the most under -used academic tools currently available to archaeologists in England .

Acknowledgements Firstly, I am most grateful to all the SMR Officers, County Archaeologists and Archaeological Officers, too numerous to mention, who so willingly gave up their time to provide the raw material for this paper, and who also provided generous hospitality to the four of us who undertook the survey! I am also most grateful to Ben Booth who gave me a technical summary of the 46 SMRs, on which information the computing section (section 3.6) is based. My thanks are also due to Ben Booth and

17

AMANDA CHADBURN

Nigel Clubb who commented responsibility alone.

on drafts of this paper. Any errors, however,

remain my

Bibliography BOOTH, B. K. W. 1988. "Site Specific Data-A logical Computing Newletter, August.

Standard

for Data Transfer",

Archaeo-

CHADBURN,A. D. B. 1988 . "A review of approaches to controlling archaeological vocabulary for data retrieval ". in Rahtz, S. P. Q., (ed .), Computer and Quantitative Methods in Archaeology 1988, International Series 446, pp. 389 - 39 7. British Archaeological Reports, Oxford. ENGLISH HERITAGE/ ROYAL COMMISSIONON HISTORICALMONUMENTS 1988. "Site Specific Data-A Standard for Data Transfer". Unpublished , widely circulated, typescript. 19th May, 1988. HISTORIC BUILDINGSAND MONUMENTSCOMMISSIONFOR ENGLAND198 5. "Sites and Monuments Records: A Standard for the Scheduling Enhancement Programme". Typescript (unpublished). 26th September 1985.

18

4 SUPERFILE - a user's view P. D. Iles* M. Truemant 4.1

History of the Lancashire SMR

The Lancashire Sites and Monuments Record was created in 1976, and since 1979 has been housed in the Lancaster University Archaeological Unit, at the University of Lancaster. Initially it comprised a comprehensive map base, together with manual files, and optical co -incidence index cards, and contained about 2000 entries. As the number of entries increased, and computer technology improved it was decided to computerise the Record, using the Central Excavation Unit's software (referred to as Version 1) and a Sirius microcomputer. In 198 7, English Heritage undertook the enhancement of County SMR's by supplying a new and more efficient software package, SUPERFILE(written by Southdata). The SMR at Lancaster was transferred to an IBM PC AT running the newly obtained SUPERFILE, and currently comprises c. 11,000 separate records, taking up about eleven and a quarter Mb of storage. The new software is considerably better and more convenient to use than version l, in particular in that it allows the use of an on -line database, with dynamic indexes , and the SMR has now been running on SUPERFILEfor over two years. In general it has produced very good and satisfactory results; there are, however, some problems with the way that SUPERFILEworks, and with how it reacts to the original structure of the Lancashire SMR (first defined in 1976). The system is still used, and fulfills most of our requirements quite well. The criticisms outlined below highlight a number of problem areas, some of which can avoided by one means or another. These problems are briefly discussed, in order that anyone moving on to SUPERFILEmay avoid them from the outset.

4.2

Superfile

SUPERFILE is a suite of programmes, written and compiled in 'C', and comprises a database manager, a sort package, a forms -based access system, a batch editor, and a unit that compresses the data into minimum storage. There is a programme called LOOK, which allows non -forms based work but is rather specialised in its use, and a programme called TAB which produces tabulated reports . As well as these, English Heritage has commissioned programs to load and unload blocks of data. The database manager runs an indexing system, in fact two parallel systems, which considerably speed up access to any specific items of data, this speed of access being a considerable advantage over many comparable systems. The data is stored as a • Lancaster University Archaeological University of Lancaster Physics Building Bailrigg Lancaster LAl 4YB t Lancaster University Archaeological University of Lancaster Physics Building Bailrigg Lancaster LAl 4YB

Unit

Unit

19

P. D. ILES AND M. TRUEMAN

a. Input :

[ 3008 ][ c.1670

X[

Bowters House ] [ Roofed

][ building

this is edited with a word processor

House:domestic ] [ PosL medieval ] [ SD ] [ 66200 ] [ 36790 ] [ ac ]

]X

to remove the sequence

XX

and superfluous b . Output

spaces to produce :

3008 Bowters Roofed building

House Hous e :domestic SD 66200 36790 ac

Post

me dieval

c .16 70

Figure 4.1 : Superfile Forms for Input and Output tagged ASCII file, and utilizes variable length fields, so that the size of the data field does not have to be declared in advance, nor is any blank space following on from that data stored (contrast this with dBase for example). The forms system (two versions, FORMSl and FORMS2 exist) is the 'front end' that the user sees almost all the time. It allows form based input and output, query and display of the information in the database. Forms are created or altered by the user, through an editor, and can display or output as much or as little of the data as the user wishes .

4.3

Problems

The problems

4.3.1

which we have with the system can be classified

into three categories

:

Minor faults

The batch loader from version 1 to SUPERFILE leaves spaces in the middle of words in text fields . These spaces have then to be edited out one by one, and record by record. This could probably be written out of the software, as it results from the way that version 1 stores the data, and how the batch loader then reads that data. In practice records are edited as they are needed, rather than spending a great deal of time on editing the entire SMR in one session. It does not appear to be possible to create forms over 80 characters wide. This limits the ability to output data. For example it is necessary to add tags to a multiple line form, and then to use a word processing package to alter the output to single lines (Fig. 4.1). The system must be reset and rebooted when the output device is changed (disc or printer). This can be overcome by using Southdata's management program, DIRECTOR (available at extra cost), which allows you to completely unload SUPERFILE without resetting the system. It is not yet possible to associate onscreen queries with sorts, in order to recover information in a desired order; nor is it possible, as yet, to produce records in a particular order from a list of record numbers entered at the keyboard. Both these features have apparently been commisioned from Southdata but nothing has been seen of them yet. 20

4.

SUPERFILE-A

4.3.2

USER'S VIEW

Medium faults

Altering a record moves it physically in the database and sortfiles are not automat ically updated. This can result in records disappearing from any subsequent print which uses a sort, if the sort in question has not been rerun after the editing session. This is a relatively common feature of databases, and is mainly due to the large processing overhead that would be required to update all the sorts for each edit that is made, and the subsequent slow operation of the database. Linking associated records is not possible with the current Lancashire SMR data structure. Some form of group record number system (eg a segmented primary record number) would be required. This would necessitate a complete rewrite of the data structure for every record. The manual, although comprehensive, is generally poorly arranged, and unhelpful. It is often necessary to flick from one section to another, and back again in order to find solutions to even simple queries. Whilst the perfect manual (for any software), remains to be written, and several revisions have been made, this is still a bad example of the species.

4.3.3

Major faults

The forms editor is very limited, and awkward to use . It is possible to bypass this editor, and alter or create forms from within your own favou r ite editor (eg . WORDSTAR or NEWWORD), but it is easy to make serious errors in the form by doing this. There is no 'online help ' available to the user from the keyboard (as is includ ed in dBase III for instance). SUPERFILE does not allow the use of extended or expanded RAM. In practice all that this means is that there is a finite limit to the size of database that you can process on a DOS based computer (most IBM and IBM clones use a DOS system) . We have reached this limit and we can no longer sort the whole of the SMR in particular fields (the TYPE field for example) . There are ways around this ; for example : l. Do not undertake

very large sorts.

Limit sorts to subsets

of the data (our current

solution). 2. Physically

subdivide

the database

into several smaller entities.

3. Buy a UNIX based operating system, eg . XENIX, (and possibly a new computer to run it) and a UNIX version of SUPERFILE(UNIX does not have the same memory limits as DOS). 4. Use another

software

package .

5. Alter the data structure to reduce the amount of processing numeric codes instead of keywords in the TYPE field .

necessary

eg. use

All of the above present their own problems, 1) is not really a long term solution ; 2) drastically increases the amount of operator time needed for work over the whole of the SMR; 3) is probably quite expensive, and UNIX is not famous for its ease of operation; 4) would entail learning to operate and tune a completely new system, which would then no longer be compatible with the large number of other SUPERFILE users; 5) would need the complete recasting of all 11,000 records, one by one . . . The most complex problem has been left to last. There are difficulties in the way that SUPERFILE deals with multiple occurrences of sets of fields in a record. Fig. 4.2 shows a typical multiple entry in version 1 format and the output that would be expected from a very simple version 1 print program. Whilst the SLEX(data

21

P. D. ILES AND M. TRUEMAN

extraction) and SUPERSORT (sorting) operations on the data were time consuming and cumbersome, the end product was a correct set of data. If this data structure is transported straight to SUPERFILE,forms are required that have multiple occurrences of fields written into them (Fig. 4.3). These multiple occurrences of the fields always appear in the order that they were input. The sets of fields within the record cannot be connected or associated together. This may be illustrated by a simple query related to finds of Roman date. Once the appropriate form has been loaded, and the model created (Fig. 4.4a) for this query ie. Roman in the 'General period' box and Finds in the 'form' box, the search can be activated . The record that appears, however, is not what is wanted (Fig. 4.4b). This record does indeed have Finds in the 'Form ' box and Roman in the 'General period' box, but they are not part of the same entry. The Roman date derives from the entry concerning the Roman road, and the presence of Finds from the entry concerning the Medieval coin. Whilst this may be conceived as a rather trivial example of this problem, it can and does cause a great deal of extra work in editing and re -editing output files to remove such unwanted records. It can also cause a great deal of confusion as to which records are actually required. It also means that it is almost impossible to utilize the TAB section of the SUPERFILEsuite. Again several solutions present themselves : 1. Commission

Southdata to adjust the programs.

2. Use another database package. 3. Adapt the SMR so that it no longer uses multiple entries. 4. Concatenate the sets of data that need to be associated, primarily the fields forming the 'Type/General Period/Specific Period/Form' section, into one new field. 5. Put up with it, and continue editing output. These 'solutions' have their own drawbacks : 1) Seems rather unlikely at present, and may well be impossible to do without great expense and much work from Southdata. 2) again means obtaining, setting up and learning a new system. 3) entails increasing the size of the SMR by a factor of about three, individually editing each record and splitting it up as necessary - this will mean the RAM limit discussed above will need to be confronted. 4) is the English Heritage solution, but has nearly as many drawbacks to it as it solves; for instance you need the data arranged in each possible permutation of the four fields, thus for a record with three entries in this set of fields, you would need a further twelve entries in the record, probably hitting the RAM problem as above. 5) is the current practice, but it is getting to the stage where the extra work involved for the Lancashire SMR may well justify another, more radical solution.

4.4

An easy solution?

Whilst the different solutions in the preceeding sections are all possible, the easiest solution to the problems (for us, not for the programmers, or for those who have to pay for it) would be to get Southdata to write patches that a) allowed us to use extended and expanded RAM, b) allowed fields to be associated with each other within a record, and c) allowed records to be associated with each other. This magic program should use the current data and utilize the same commands as before. Whether this is possible is another matter entirely .

22

4.

SUPERFILE-A

USER'S VIEW

a. Portion of Version 1 Data Structure from Lanes SMR 35 06 07 08 11 13 31 33

Primary Record Numbe r District Parish Nationa l Grid Re ference Site name/main features Type/general period/specific period/form Created - updated by/mon th/ ye ar OS 1:10,000 or 1:10,560 quarter sheet

b. Sample of SMR entry in version 35 06 07 08 11 13 13 13 31 33

4720 Wyre Bleasdale SD 582434 Water y Gate/Medieval coins found on site Metal:coin/Medieval/-/Finds; Metal:coinhoard/Med ie va l/- / Fin ds; Road/Roman/-/Other structure MF/ 11 /1 9 8 5 SD 54 NE

1 format

of Roman road

Figure 4.2:

[ SD 54 NE ) [ Bleasdale PRN [ 4720) Parish [ Wyre NGR [ SD) [ 58200] [ 43400) [ District [ Watery Site Name Gate [ Medieval Site Features coins found on site of Roman road Type Metal:coin l Form [ General Period Medieval ( Spec Period Type Metal:coinhoard l Form [ General Period Med ieval [ Spec Period Type Road l Form [ General Period Roman [ Spec Period Type l Form [ General Period Spec Period [ Type l Form [ General Period [ Spec Period Type l Form [ General Period Spec Period [ Type l Form [ [ General Period Spec Period [ MF l [ 11 l [ 1985] Updated - Created

Figure 4.3:

23

)

l Finds -

Finds

Other -

structure

l l l l l ) )

l l l l l l l

P. D. ILES AND M. TRUEMAN

a. Form to Search Database Parish [ PRN [ l [ District NGR [ 1 [ l f l [ Site Name Site Features Type Roman General Period Type General Period Type General Period Type General Period Type General Period Type General Period Type General Period Updated - Created [ MF' l [ 11 l ( 1 985]

I

J Focm Spee Period I Form Spee Period I Form Spee Pc r L od I Form Spee l,e r i od I Form Spee Period I 1:-'orm Sp e Period I Form Spee Perioo

Finds

l

l

l I l I I l I I 1

b) Result produced from database Parish l Bleasdale PRN [ 4 720] [ SO 54 NE] District l Wyre NGR [ SD] [ 58200] [ 43400] [ Site Name [ Watery Gate Site Features [ Medieval coins found on site of Roman ro�d Metal:coin J l•'orm l Type General Period Medieval Spee Pe i�d [ I l·'orm Metal :coinhoard Type Spee Period I Medieval General Period ] Form I Road Type Spee Peri:;d I Roman General Period I Form I Type Spee Period General Period Type I Form Spee Perioo General Period Type I l·'orm Spee Period General Period Type I Form Spee Period General Period Updated - Created [ MF' J [ 11] [ 1985 I

Figure 4.4:

24

l•'inds l•'inds OLher sLrucLure -

J )

l 1 I l l I l

SUPERFILE - A USER'S VIEW

4.

Ribble Valley District Roman SD 60

Sites

Sites Sites With Finds

Lancashire

+ EE

Sites and

_fB'I.. Ill ·. : ·le ·...

·.'at · ·

+-f

..

. •·.

SD 30

,---,~---= ,

0 SD 60

SD 70

Figure 4. 5 : Example plot

25

5 Km

P. D. ILES AND M. TRUEMAN

4.4. l

Current thinking on other solutions

Given the problems above, the size of the current database, the use of Oracle by English Heritage and RCHM and the availability (with in-house support) of Ingres, there appears to be a better solution for us. An upgrade to a relational database associated with a geographic information system (another complication, outside the scope of this article) held either on the University of Lancaster mainframe, or a powerful mini system such as a SUN workstation, is now being considered.

4.5

Conclusion

SUPERFILEis undoubtedly a good, flexible, and convenient database package, and has successfully run the Lancashire SMR for some time, and continues to do so. It should be emph~sised that most of the problems described above result from how various 'features' of SUPERFILEreact to the the structure of the data within this particular Sites and Monuments Record. These problems are related in particular to the multiple occurrences of fields used in the Lancashire SMR and its sheer size. Anyone at present considering the adoption of SUPERFILEwould be well advised to avoid the use of such multiple entries, and to utilize some form of group record number and subsidiary record number method. SUPERFILEdoes not provide the perfect SMR system, and given the very wide variety of tasks required of an SMR, it is unlikely that any such perfect system will ever be made available. Nevertheless, the use of SUPERFILEas the main means of data input, indexing, interrogation, and providing formatted output for the Lancashire SMR over the last two years, has greatly increased both the general efficiency and use to which the SMR has been put during this period. In addition, the resources released by the efficient management of this record, as a direct result of the use of SUPERFILE,has allowed a considerable increase in the direct output derived from the SMR, for example, planning control including assessments for local and district plans), and more generalized survey and assessment including the Monuments Protection Programme, as well as the development of new and associated facilities, such as automatic computer generation of distribution maps.

26

5 Using dBase for county SMRs: the Humberside experience John Wood* 5. l

Introduction

The establishment of archaeological data -banks at local or regional level has been an important development in British archaeology over the last 15 years. Unfortu nately, those who set up or manage these Sites and Monuments Records (SMRs) often encounter problems in developing and justifying the services they can provide. Frequently working in isolation, subject to political, financial, or other pressures, time and money may be wasted seeking solutions to problems already solved elsewhere. Alternatively, inappropriate strategies and structures may be adopted, leading to longer -term difficulties, and preventing SMRs from reaching their full potential. This paper attempts to explain one particular approach to establishing a County SMR, and why and how dBase software was used to computerise it. It also takes a wider look at the use of dBase for SMR applications . Particular emphasis is placed on the importance of designing systems around their expected user requirements and the need to consider the practical constraints which may be placed on their development .

5 .2

Why create an SMR?

On arrival in Humberside in September 1984, the writer's brief was to establish for the first time an organised Sites and Monuments Record for the County . There was a requirement from both the Historic Buildings and Monuments Commission (HBMC), who were providing initial 'pump -priming' funding for the project, and the County Council, who were expected to take it on, to demonstrate results as soon as possible. However, the requirements of the two organisations were different . As with other County SMRs, HBMCwanted the record to form part of a national network of county based records which could , among other things, help them assess and identify sites and monuments of national importance for protection by Scheduling. Some degree of compatibility with other SMRs was therefore essential. The County Council on the other hand was not fully convinced it wanted an SMR- after all, it had managed without one up to that point. Local Authorities exist to provide services to the community, and for them the SMRwould have to justify itself in public service terms. Within the County Council structure, the record was to be based in the Humberside Archaeology Unit, which had also previously managed without an SMR. The main requirement here was for background and comparative information which could be used in setting up field projects and research priorities for the county. To be successful, therefore, the SMRwould have to provide as wide a variety of ser vices to as many people as possible in the shortest possible time. It would also have to market these services to build up support for continuing the project. Resources and equipment (such as a computer) would need to be attracted to the project and "' Department of Archaeology University of York Micklegate House Mickle gate York

27

JOHN WOOD

justified against competing demands. Spending months or even years laboriously preparing, checking, and entering data to a machine in the vague hope of future reward was therefore out of the question. Before it could offer a service however any system adopted would have to be 'fully retrievable for all reasonable purposes of enquiry' (Baker 1983, p.5 5), revisable with new evidence, and capable of development and change. Development plans would also have to be realistic and take into account the very limited resources available. Fortunately a good deal of preliminary gathering of material from the Ordnance Survey, Yorkshire Archaeological Society and other sources had taken place, and Loughlin and Miller had published their Inventory of Archaeological Sites in Humberside in 1979 (Loughlin & Miller 1979) .

5 .3

Who will use it?

Many people seem to believe that Sites and Monuments Records exist only for de velopment control and heritage management purposes. Some will also recognise their potential for defining fieldwork priorities. However an SMR can and should aim for a much wider role than this. It should b e available as a resource for students, researchers, and teachers at all levels ; archaeology, local history, and civic societies ; authors of academic and popular books, leaflets, and other material; farmers and landowners; and other enquirers. An SMR should also be able to promote archaeology positively and identify new possible uses and services. Of course, the SMR officer cannot be expected to allow people to use the information contained in the record to damage or destroy sites, or to steal the property of landowners and occupiers. Balancing the needs of conservation and security with those of public information is therefore an important consideration. The structure of the SMR should make it possible for the controlling officer to make a professional judgement regarding access to information in individual cases.

5 .4

The strategy

adopted

At the very least, an SMR should be able to provide a basic retrievable index to archaeological resources for research, conservation and management, education and presentation. It should be able to tell us quickly what there is, where it is, and what it is currently thought to be. Provided the information held is structured properly to distinguish primary and secondary sources and make the updating process explicit, 'What and where' provides at least a basis for professional judgement, as well as a point of departure for research. No SMR is ever 'complete', and no matter how so phisticated it is, a critical approach is always necessary in assessing the information held. A planner, road engineer, or building surveyor needs to know whether there is any archaeology at a given place, and what its implications are likely to be for his work. A researcher on the other hand may ask where are there examples of a particular site type or status. As the record develops, higher levels of evaluation and interpretation can and should of course be included, but they ultimately depend on this base level of information. A basic, county -wide database was therefore created from the outset and its scope gradually increased, instead of working slowly across the area in detail. At the same time, a high -profile approach was adopted from the beginning. Every opportunity was taken to advertise and develop th e record' s potential within the County Council, to the relevant District Councils, and more widely. Contact was made with architects, road engineers, building surveyors, estates and valuation staff, education advisers, planners, countryside, minerals, and forestry sections. Up-to date lists of sites and monuments, followed by 'constraint' (consultation) maps, were sent to the nine Districts and meetings arranged with re levant officers . Short reports 28

5.

USING DBASE FOR COUNTY

SMRS: THE HUMBERSIDE

EXPERIENCE

appeared in the County's Education Bulletin, the newsletter of the Rural Community Council, and the local press. Leaflets went out to Adult Education groups, local libraries, museums and societies. Close collaboration was developed with the local Farming and Wildlife Advisory Group. An early decision was made to create an essentially paper record, using a computer, if possible, to do what it is best at : i.e. to provide a retrieval system. There can be no particular value at present in attempting to transfer all notes and references held in SMRs to the computer. Generally speaking, reports need to be individually written for whatever purpose they are needed, and cannot be satisfactorily assembled as kits from standard computerised blocks of free text. Similarly for those enquirers who want detailed information, the photocopier is still in 1989 a much more cost effective piece of technology. The day of the fully -computerised SMR will no doubt come eventually, but when it does OCR (Optical Character Recognition) equipment should make data entry easy and cheap. Typing in unnecessary free text wastes time and money, and takes up valuable computer space better used for something else. Besides, a paper record is usually necessary in practice for storing photographs, drawings, field -notes, old planning applications, and correspondence. Later, as the computerised record develops, it may prove useful to include in it a short summary of the current state of knowledge on each site to help deal quickly with straightforward enquiries. Other ways of saving time and money were also found. Instead of creating index cards or data input sheets, the existing copies of Ordnance Survey, Yorkshire Archaeological Society, Hull Museum, and other records were roughly stapled together in the first instance and given Primary Record Numbers (PRNs) . These were then stored in sequence and used as they were for computer data entry . Meanwhile, the references were transferred gradually to standard record sheets kept in box files. Simple parish lists, containing the PRN, identifiable site name and status (Scheduled or not) of every site in each parish, provided a means of access, and a Record Manual was produced in a form which could be easily updated, designed as much for the user of the system as the compiler. Keywords and categories were adapted from those in use in Bedfordshire SMR, since having been designed originally for an optical co -incidence card retrieval system they were straightforward and unstructured . A Community Programme team funded by the Manpower Services Commission began work on a set of basic 1: 10,000 dye line maps. Meanwhile, demonstration of the SMR's potential persuaded senior County Council managers to transfer to the project a twin -floppy drive 256k Sirius from another section.

5.5

dBase: Why and How

After some experimentation with other software, dBase II was chosen mainly because it worked, and as it was already in use in the County Council, it was free and there were people around who knew how to use it. However, its potential was immediately obvious. A retrievable data -bank could be easily established and interrogated in a variety of ways. It was flexible and the data structure and format could be altered and developed when necessary. It could be used interactively, or simple programs could be written for specific applications. Being a widely used, standard, commercial software package, it would also be easy to upgrade and if necessary transfer the SMR to new software when it became available without typing it all in again. Data could also be transferred to other records, since the flexibility of the package makes it possible to modify data structures and words used to fit the receiving system. One person's 'matte and bailey' can easily become another's 'fortification: castle'. Because it has structured records it is also possible to transfer information via ASCII to other types of record system (such as those based on Superfile). It 29

JOHN WOOD

is interesting to note that this facility for data transfer has since grown, and both the North Yorkshire SMR system and the English National Monuments Record (both mainframe-based) can now interface with dBase III. Most important tqough at this stage was the fact that dBase could be used to create a demonstrable, working computerised SMR extremely fast and cheaply. Using the collected notes and references a county -wide database was built up very quickly from the parish lists using global replaces where possible. A simple file was set up for the first parish, containing the same basic details as the list: Primary Record or reference Number (PRN), an identifiable (unique if possible) site name, and its status (Scheduled Monument or not). The parish name was then entered using a global replace, the file structure copied to the next parish and the process repeated. When all the parishes in a district had been entered, the parish files were combined to create one for each district and the global replace used to enter the district name and pre -1974 county . This method was not only much quicker than entering everything by hand, but also minimised typing errors. A decision was made to exclude from the initial data input phase information which could logically be deduced from material already entered : this could be done later if required, using either the global replace function or simple specially -written applications programs. The file structure was then enlarged to include other fields, such as grid references, site types, and periods represented, and the relevant details entered. The limitations of the twin drive computer meant that it was necessary to store each district on a separate disk-in effect 9 separate mini -SMRs arranged to the same format. Indeed, one of the districts (East Yorkshire) is so large and rich in archaeology that it soon had to be split into two. Despite these hardware problems, dBase proved to have all the required virtues of flexibility, simplicity and user -friendliness, being easy to use interactively by the writer or other staff without much need for high levels of computer literacy. Listed Buildings, SSSis, Ancient Woodlands, Forestry Commission land, and other retrieval categories were added. Within a year we had a basic, retrievable SMR for the whole county, although because of the fragmented nature of the record County -wide searches were cumbersome . The next stage was the acquisition of a (second -hand) 10Mb hard disk Sirius, which allowed us to put the whole SMR into one file. Simple programming in dBase turned out to be surprisingly straightforward. Pro ducing computer -generated parish lists was no problem, and a special program was devised to check and standardise the National Grid map references, which had been entered in a variety of ways. The program also used the grid reference to identify the relevant 1: 10,000 Ordnance Survey maps and enter these automatically in a new 'map' field. From this a map list was generated to check and update the maps themselves. As the computer database grew, mistakes in data entry inevitably appeared despite (and sometimes because of) the use of global replaces. These are powerful and dangerous commands, and care has to be taken so that only those records which require changing are affected. Mistakes produce absurd (and often hilarious) results, which can take some time to correct! However it was usually possible to formulate checking routines and carefully designed searches to bring these to light. In autumn 1987, when the SMR had outgrown the 10Mb Sirius, HBMC made it possible to transfer to a new 40Mb IBM(AT) compatible NEC computer running dBase III+. This had a number of advantages, not least that it was now possible to add a 'memo' field containing a free text summary to each record. It also allowed 15 files to be open at once instead of only two as with dBase II, creating the opportunity to develop a network of relational databases. A relational database system is particularly appropriate for SMR applications. Sites and Monuments Records need to cope with a variety of enquiries, and different types of site may require different categories or levels of information. Some retrieval fields 30

5.

USING DBASE FOR COUNTY SMRS: THE HUMBERSIDE EXPERIENCE

may only be required for Scheduled Monuments or Listed Buildings, for example, so sub -files can be created for these, linked by the PRN to the main database. These sub -files are usually small and quick to search, and reduce the need to take up unnecessary space on the computer by storing unwanted blanks. Cross -references to other records can also be provided where applicable-in Humberside, these include slides in the Unit's collection, HBMC, Yorkshire Archaeological Society and National Archaeological Record reference numbers, the register of cancelled numbers where new PRNs have been allocated, and details of excavation archives. The relational arrangement can be taken further by networking several machines. Apart from providing extra terminals, where the record is housed within a field unit it should be possible to make direct links if desired between SMR, post -excavation, finds, drawing office and other databases. Other links could be established with local authority planning, property and estates records, with defined levels of access. With a MODEM, a number of similar SMRs could readily form a regional network. Many people criticise dBase for its fixed -length fields, but this is not a serious problem if the system has been designed to make use of the relational facility. The length of retrieval fields is of course defined by the user, and can be altered if necessary, but the existence of fixed length fields can have a positive advantage in providing an informal structure to the data. Storage space on computers is rapidly becoming much cheaper, and a few stored blanks can usually be accommodated if space is not wasted on repeating fields, unnecessary free text and over -complex data structures. Similarly, abbreviations and coding have been avoided in the Humberside SMRwhere possible, as the advantages of a simple, user -and operator -friendly system were considered to far outweigh any possible benefits from gained computer storage space . dBase IV, which has recently become available, has a facility to translate codes . Database linking is also improved and there are better report and query generators. The Assist function in dBase III+ has been replaced with a 'Control Centre' which covers most things, reducing the need for a programming language. There is also an on -line Help facility. Query by example (QBE) and the structured query language SQL are available, and the dBase programming language is still there (in improved form) if required. For anyone still using earlier versions it should be well worth the upgrade, although it seems slow to work in certain configurations. The Humberside SMR continues to grow and develop, serving not only the planning process but also the wider community. Mary Lakin, who took over as Sites and Monuments Officer in January 1988, strengthened the data structure and greatly improved the map record, and her successor Ed Dennison will no doubt make further changes. This is as it should be. SMRs are dynamic in nature and must change and develop continually to meet the demands placed upon them. In due course, dBase itself will be obsolete and all the records will be transferred to a new format, but for the moment it still has much to offer for developing Sites and Monuments Records.

Bibliography ACAO 1978. "A Guide to the Establishment

of Sites and Monuments

BAKER,D. 1983. Living with the Past: The Historic Environment. BURROW,I., (ed.) 1985. County Archaeological LOUGHLIN,N. & K. MILLER1979. "An Inventory side". Wooo, J. 1984 -7. "Reference notes, information Humberside SMR". (unpublished). 31

Records".

Batsford.

Records: Progress and Potential.

of Archaeological

Sites in Humber -

sheets and record manuals for the

JOHN WOOD

MAIN FILE: location, description, status form, condition, land classification is other information held? (YIN)

'V

MEMO FIELD: short summary of current knowledge

I link provided by Primary Record Number

l Slide collection index database

l I Index of cancelled Nos. I

subfiles containing cross-references and special fields for particular site types not required elsewhere

I National Monument Record: NAR (OS) No(s). and details

~

' /

Listed Buildings: Ref.Nos and other details specific to Listed Buildings (Grade held in main file)

'Loutlin & Miller' quoted source) page references, where applicable

(= muc

Index to Pottery and Finds Archives

~

,

Scheduled Ancient Monuments: County No(s). and other details speafic to S.A.M.s

Yorkshire Archaeological Society Inventory numbers

/

~

'

,I

~

' /

Excavation Archives: index to Site Codes, location of material, excavator's name, contact addresses etc.

Aerial Photographs index

(one Humberside PRN may include several NAR Nos./ YAS Nos./ etc., or several PRNs may refer to one, since the information may not have been divided in the same way. The computer finds all the relevant matches and displays them . Searches can be done from indexed sub -files to the main file or vice -versa.) Any number of linked files are theoretically possible. dBase III+ allows 1 S to be open at once, which is probably plenty for most purposes ; dBase IV allows 99 files to be open simultaneously.

Figure 5.1: Relational structure

32

of the Humberside

SMR

6 The Greater London Sites and Monuments case study

Record - a

Hugh Jones* 6.1

Introduction

CAA conferences are concerned with the application of computer technology to archaeology. However, the computing problems and opportunities which are presented by the Greater London Sites and Monuments Record can only be properly understood when placed in the context in which the record has developed and is used. For this reason this paper falls into three main parts. The first (sections 6.2 - 6.5) describes the organisation of archaeology and conservation in Greater London, and the way in which this has affected the development of the record. The second part (sections 6.6 - 6.9) goes into some detail about the hardware and softwa r e currently used, and the advantages and disadvantages which they bring. Finally (sections 6.10 - 6.11 ), the paper explores directions which might be taken in the future.

6.2

History of the SMR

Greater London was one of the last English counties to start developing a county -wide sites and monuments record. It was also one of the first to consider keeping information on archaeology and historic buildings on the same record. This is because the initiative for setting it up came from the Greater London Council, more particularly the GLC's Historic Buildings Division, where archaeology was very much subsidiary to a variety of work relating to standing historic buildings. This initiative came in the early l 980s, and happened when and where it did for a number of reasons: the GLC wished to have more sophisticated ways of handling the information it both used and generated in its work on historic buildings ; there was an enormous amount of archaeological work, both past and present, whose results needed to be systematically available to a wide range of users; and it was recognised that almost every English county had or was developing a computerised sites and monuments record while London did not possess one. In addition, it was felt that the very real fear of abolition which was then looming meant that there may not long be an opportunity to establish such a record London -wide. In 1982 the GLC funded a survey by the Museum of London to establish the range of source material for the archaeological part of the record. At the same time the GLC looked at how it wanted to manage its own information relating to historic buildings, and assessed the computing facilities available within its own Central Computer Services. In 1983, the GLC employed an SMR Officer to carry out the detailed design of the record, to set up and co -ordinate the team to work on it, and to oversee the creation of computer systems for it. Compilation started in 1984, and was, and still is, carried out by a team whose members are employed variously by the Museum of London, the Passmore Edwards Museum, and the Borough of Kingston * English Heritage HBMC Chesham House 30 Warwick Street London WIR 6AB

33

HUGH JONES

on Thames. Most of the funding for these posts came from the GLC, and to date the same arrangements have been maintained by English Heritage. On the abolition of the GLC in 1986, English Heritage took over the majority of the functions of the Historic Buildings Division of the GLC, which, with minor changes, became the London Division of English Heritage. The effects of this change are important to the future of the computing strategy, as will be shown later.

6.3

Conservation in London

One of the major determinants of the structure of the record, and the computer sys tem which supports it, has been the range of interests which the SMRhas to cater for. In London, the GLC had, and English Heritage now has, powers to direct local planning authorities in the granting or refusal of listed building consent. Elsewhere in the country, these powers are exercised solely by the Department of the Environment. There are 33 local planning authorities in Greater London (31 boroughs plus the Cities of London and Westminster), and these should get appropriate advice on the archaeological implications of planning applications from the Museum of London and the Passmore Edwards Museum, which use the SMR for this purpose. There is also the London Docklands Development Area in which planning rules are somewhat different from those elsewhere. London has, in the DUA and the DGLA, what are popularly believed to be the two largest archaeological field units in Europe, as well as the Passmore Edwards Museum active in North -East London, Kent Archaeological Rescue Unit in South-East London, and numerous museums and local societies. There is also the need in London, as elsewhere, to consider the interests of the major national bodies in the field, the RCHM and English Heritage, and the interests of academic researchers, and other public and commercial users. One way of making the SMR responsive to the needs of this wide constituency has been to base the members of the SMR team in a number of the organisations both providing and using data. Another has been the existence, from the beginning of the project, of an Advisory Group, on which representatives of many organisations can monitor progress and provide an input into future policy making.

6.4

The structure of the record

The information needs of the different interests identified above are many and various, but they had to be reconciled in a single working system. This required decisions on a number of points, and decisions according to a variety of criteria: some philosophical, some practical, some technical, and some no doubt entirely arbitrary. One of the first decisions to be taken was that it was appropriate to put both historic buildings and archaeological items on the same record. This was decided partly because it is often an accident of history whether something has survived to the present day as a standing structure, and partly because there can often be very real uncertainty as to whether something should be considered as an archaeological site or a historic building, The Tower of London for example, or Hampton Court Palace, where the debris resulting from the recent fire was removed archaeologically; there will also be times when the management interests of archaeology and historic buildings will overlap. Another decision to be taken was the level of detail which it was appropriate to allocate to a single record. The solution adopted was to take the idea developed in the North Yorkshire system, whereby each record has an equal status in the database, but groups of individual records may be related to one another as primary records, component records of a primary, or sub -components. This means that there is great flexibility, so that a single record can refer to a complex archaeological site at one extreme, or a single listed bollard at the other. 34

6.

THE GREATER LONDON SITES AND MONUMENTS RECORD-A

CASE STUDY

It had to be decided whether free text retrieval was desirable, and if so, if it was achievable. It also had to be decided what classes of information to record, and whether there would be a substantial storage overhead from having a large number of fields defined, many of which might be empty in any particular record. The choice was to record archaeological sites in a fairly standard SMR-type way, but include further fields particularly relating to historic buildings, and yet others reflecting the particular management responsibilities then enjoyed by the GLC. A decision was needed on what terminology to use, whether to use abbreviations or whole words, what to validate, how many indexes to define, and so on and so forth. However, the main point is that computer -based record systems end up looking and performing how they do for a variety of reasons. Some of those are indeed to do with the hardware and software used, but technical considerations are only part of the picture. The opportunities or limitations provided by any particular computing environment can often have less of an impact on the effectiveness of an application than other, non -technical factors, such as resources, objectives - and that includes not just the objectives set at the beginning of an exercise, but the ability to review those objectives and amend them accordingly - the corporate culture of the organisation running a record, the adequacy of communication with other relevant organisations, and the closeness of the match between the skills possessed by those working on the record and the skills needed to carry out the work effectively . For these reasons, while it is important to obtain the best possible computer system for any particular application, it is at least as important to take a holistic view of the problem to be solved, and ensure that whatever computing solution is chosen will form a coherent whole with the other factors which must come together to make an application succeed.

6.5

The contents of the record

In the circumstances in which the Greater London SMRwas set up, it was decided that it was necessary to develop a large record. It is believed that the current phase of compilation, which began late in 1984 and is expected to finish in 1991, will produce something in the order of 65,000 records, of which rather under half will relate to archaeology. This 'basic record', as it is called, consists of all archaeological sites and isolated finds about which information has been available from reasonably accessible sources, and all statutorily listed buildings. The need to add to this basic record in the medium term has been reviewed, and the resources available to do this have been assessed, and it is believed that a good case can be made for doing further research into the less accessible or productive archaeological sources, and adding information on historic parks and gardens. Some people feel that the SMR should become a much broader -based record dealing with all aspects of the past of London, but the resource implications of this are enormous for a record which is already many times the size of an average county SMR. The potential size of each individual record on the system also had to be quite large. In order that the database could truly hold both archaeological and buildings data together, there is a single record structure to cater for all data. There is thus a significant amount of redundant space in every record. As there was a wide range of user interests identified from the start, there is a large number of fields defined.

6.6

The current computer service

These two things, the number of records, and the potential size of each record, meant that the computer system had to be quite large and powerful. In addition, the need for direct access to the record from a variety of locations around London meant that

35

HUGH JONES

there had to be good communications facilities. In these circumstances, a PC-based record was not a viable option. The GLC therefore chose, not surprisingly, to use their own in-house mainframe computer service for the record. This had the advantage that hardware and software were already in place, communications across a wide area were well established, and the cost of developing and running the system was internal to the one organisation. There have been minor changes to the SMR computer system since development started, but these have been evolutionary. The computer system is run by the Central Computer Services, or CCS, of what used to be the GLC and then became the London Residuary Body . In 1988 Hoskyns plc acquired CCS as part of the process of disposal of all the former GLC assets. Since the abolition of the GLC, CCS' biggest customer has been the Inner London Education Authority, which has a number of applications including payroll, supplies, building maintenance, property register and management information systems. Other major users include the London Fire and Civil Defence Authority, the London Research Centre, and a number of the London boroughs. Since abolition, English Heritage has paid for the computer service on a 'cost only' basis - that is to say that the Residuary Body only covered its costs and did not make a profit. Hoskyns have agreed to increase their prices only by the rate of inflation for three years, and will introduce profitability by cutting operating costs. CCS are responsible for provision and maintenance of hardware and software for the SMR, and have day -to -day responsibility for all technical aspects of the record such as ensuring the availability of the computer service, security, data prep., and programming (apart from a certain amount of updating and report generation carried out by SMR staff). English Heritage retains, through SMR staff, all responsibility for the data in the database except where it has been corrupted by faulty CCS software, and for controlling access either to edit or to read the database.

6. 7

Hardware and software

The computer itself, which is currently at County Hall on the South Bank of the Thames opposite the Houses of Parliament, is an IBM 3090 running the MVS/XA operating system, which controls the internal processes of the computer. Between the operating system and the application is a facility called TSO (which stands for Time Sharing Option) - this splits up the available computer time between all the users. The database used for the SMR is called ADABAS, which is produced by a company called Software AG. This was chosen as the more appropriate of the two database packages available on the County Hall mainframe. ADABAS is sometimes called a 'semi-relational' database - each database can be split into up to 25 5 files, but the information in each of those files need not be related to the other 254 in any way. The mainframe at County Hall has only two databases on it, and the SMR uses only two files out of the 2 5 5 available on one of those databases, one for the main descriptive material and the other for bibliographic and archive reference data, so in fact it is essentially stored in two intensively indexed flat files. Each ADABAS file can hold up to 16. 7 million records, and each record is up to 500 fields long. ADABAS has an associated programming language called NATURAL, claimed by the manufacturers to be a fourth -generation language. NATURALis used by SMR staff to produce customised reports from the record. ADABAS has a number of features worth commenting on. As with many systems, screens can be set up to suit the needs of the user, and information can be retrieved using either a menu-driven or a command-driven system. The menu-driven system is easier to use but slower and less flexible, while the command -driven system is more suitable for experienced users. A maximum field size is defined for ADABAS fields, but the database only stores what data is put in, and not the blank space in

36

6.

THE GREATER LONDON SITES AND MONUMENTS

RECORD-A

CASE STUDY

vacant or half -empty fields. Currently, only about 20% of the total available space in the SMR database is actually used, so this facility makes for much more economical storage than would otherwise be possible. It also allows records to contain free text, although that free text is not indexed or otherwise retrievable in any way. Indexed fields are called descriptors, and it is possible to set up descriptors at any time, after data has been entered as well as before. It is also possible to set up indexes on a specific combination of fields or on a specific part of a single field: these are called super -descriptors and sub -descriptors respectively. It is possible to search on fields which are not indexed, but this requires a sequential search through the whole database, and is very expensive, not just in terms of machine time and disk accesses, but also in cash terms. Certain categories of data may recur an unpredictable number of times in any particular record. ADABAS caters for this through multiple fields or periodic groups. These are single fields, or combinations of them, which may be so defined that they will be present in any record the number of times that there is data to go in them. They must however be used with caution, as it is not possible to sort records according to data categories defined as multiple fields or periodic groups .

6.8

Computing

blind alleys

Before stating the advantages and disadvantages of the current computing arrange ments, and by that is meant not just the technical but also the organisational factors, two computing blind alleys should be mentioned which have been explored in the past. The first is STAIRS, a free text retrieval system. This was the second of the databases available from CCS whan the SMR was set up. There was a pilot study to assess the value of STAIRS using some of the text from the statutory lists of historic buildings. The particular problem which ruled it out was that it was designed for very static data, and therefore did not possess any in -built editing facility. In order to amend any typing errors, or make any of the changes or additions which happen to the lists from time to time, it was necessary to download the whole file into the somewhat primitive text editor available on the mainframe, make the necessary changes, and reload the file. This was not a workable system. The second piece of software which has been tried and rejected is called FOCUS. This is the report generator element of another database system, which is available without the main part of the FOCUS package and can be interfaced to other databases. We have experimented with it because it has 2 modes of operation, menu -driven and programmed, whereas NATURAL,which is currently used, has no menu -driven report generation facility. The FOCUS menu -driven system was indeed easier to use than the programmed version, but was also very much less powerful, so much so that it could not perform the required tasks. Rather than re -train the relevant staff to program in FOCUS, we have elected to continue using NATURAL.

6.9

Advantages

and disadvantages

of the current computer

system

The following are seen as the advantages of the current system. It is designed for London -wide use, and for large records. It is run by a large computing department which can afford to have specialists in different aspects of hardware and software. ADABAS is a mature product and is not subject to frequent bugs, and it can mix structured data and free text and provide a range of facilities such as data validation and sophisticated indexing, including the ability to create new indexes on an existing database. New fields can also be defined and added at any point. System security is also well established - as well as regular back -up procedures, access to databases is properly controlled according to need.

37

HUGH JONES

The disadvantages are, firstly, that the system is expensive. Secondly, its future is uncertain under new management. Thirdly, the level of service cannot be guaranteed. Users are obliged to accept the advice of CCS on technical matters, and work has to be paid for even when it is abortive-the arrangement offered is that payment is for time expended, not results achieved. The SMRdatabase is perhaps more complicated than, and certainly different in kind from, the majority of databases managed by CCS, and programs do not always seem to be adequately tested before they are run. Where these programs make significant changes to the database, more harm than good can result, causing extra work for SMRstaff who have to make good the damage. Every user, including the SMR, is subject to problems affecting the whole mainframe system - there are many ways in which such a system can be brought down, and from time to time the system is unavailable for use, and we have no control over when it comes back up. On a more specific level, there are genuine problems caused by the inability of ADABASto sort on multiple fields, and there is not sufficient flexibility in the formatting of printed reports to allow them to look how we would wish.

6.10

The future direction of Greater London SMR computing

The disadvantages of the current computer system are now believed to outweigh the advantages. There is also a changed standpoint since the SMR has been run by English Heritage. Under the GLC, there was a need to set up a system, and get it running, quickly, before abolition. Cost was perhaps not as important a consideration as it is now, as the computer service was free at the point of use. An awareness of the true costs of mainframe computing might make many archaeological users of mainframe computers think again about the appropriateness of that technology. In English Heritage, unlike the GLC, conservation is the primary function, and the organisation is already involved in a considerable way in archaeological computing, in such areas as the Central Excavation Unit and the Ancient Monuments Laboratory, as well as the initiatives in SMRcomputing, including grant -aid, as well as the support of the so -called 'Version I' software and Superfile. In this context, English Heritage is considering the computing strategies of three organisations: English Heritage, the RCHMand the Museum of London. In particular, we are looking to the agreement on computing standards reached by English Heritage and the RCHMto guide the future development of the SMR.The objective is therefore to develop a replacement computer system in-house for the SMR, based on the rela tional database ORACLE.It is possible, with many SMRs now outgrowing their current computing environment, PCs getting more powerful and cheaper, and ORACLEnow available on them, that this may become a new standard in SMRs. It is not easy in the public sector these days to specify computer hardware for more substantial systems, due to the Common Market GAIT regulations which require such purchases to go out to tender. English Heritage will look to follow the RCHM's lead in getting DEC hardware, with either the VMS operating system or a version of Unix, but failing that, there are many other hardware platforms for running Unix. The SMRwill wish to retain as many of the advantages as possible of the current system, while removing as many of the problems as possible, and in particular we consider it essential to allow direct access from different locations around London. Another important development will see the implementation of easy-to -use enquiry systems to cater for the most common types of query.

6.11

Mapping

In conclusion, some brief ideas on potential developments in mapping. The items recorded in the SMR are currently marked on film overlays attached to 1:1250 or

38

6.

THE GREATER LONDON SITES AND MONUMENTS

RECORD-A

CASE STUDY

1:2500 scale O.S. maps. This is adequate for the moment, but no more, and in the medium term there will be a need to develop a new system, which it is hoped would be computerised. Two systems are immediate possibilities. One is available from CCS at County Hall and is based on the Ordnance Survey's own computerised map base-London will be one of the first areas of the country for which the entire map base will be available. At the moment, we fear that the cost and level of service which, based on our experience with the SMR, we might expect, will probably rule out this option, but we are watching its development with interest. The other possibility is the system based on raster -scanned maps used in the English Heritage Records Office in Fortress House to record the location of scheduled ancient monuments. At present the map coverage is only available at 1: 10000, but such a system with large scale maps could form the basis of a solution for the SMR. In the long term, there is no doubt that SMRs will make use of fully -fledged Geographic Information Systems, preferably in conjunction with many other categories of data, but it may . be some time before such systems are fully functioning in a form which is affordable to most archaeological users.

39

40

7 Sites and Monuments records in historic towns - problems in the development of computer records for urban area s N. A. R. Lang * 7 .1

Introduction

The identification and preservation of urban archaeological deposits has been of growing importance through the late 1970s and 1980s, as re -development pressure has increased the threat to extant deposits . Despite great advances in the develop ment of archaeological record systems this decade, the need for a detailed machine based record of historic towns has received limited attention. T he purpose of this paper is to highlight some of the problems in establishing such a computerised record, and some of the most pressing needs in database design and development. There are several agencies which consistently record data from historic towns. These include the National Archaeological Record (NAR), the Records Office of English Heritage (EH), and the County Sites and Monuments Records (SMRs). It is the latter recording system which would normally provide the most comprehensive dataset for the archaeological and historical development of a town, though many of the problems faced by SMRs are also common to these other recording bodies. Most Sites and Monuments Records have either completed or are approaching the final phases of recasting their paper records to computer. The process of com puterisation has been extensively encouraged and supported by English Heritage. This support has included the sponsorship of software for SMRs (initially 'Version l' developed for the Central Excavation Unit, now replaced by Southdata's 'Superfile' package), and telephone support. The involvement of English Heritage has had a considerable influence over the data structure and content used by the majority of the County SMRs. Many have adopted the 'AM107' structure (Fig. 7.1) u sect to record the schedule of ancient monuments, with minor modifications. It is generally acknowledged that the AM107 structure is adequate for most planning needs related to rural archaeological sites. The data fields used can also answer the majority of public and 'low level' academic enquiries. The potential use of SMR's to pursue wider goals may, however, be limited by close adherence to this structure. The purpose of this paper is to highlight the problems and possibilities inherent in recording one specific site environment, the urban area. It is contended that most SMRs are using a data structure which was developed primarily with the attributes of field monuments in mind. The limitations this structure presents for urban records has limited the content and discouraged widespread use of these databases, particularly in research. For the purposes of this paper, 'urban deposits' refer to the total built historic environment (both above and below ground) in areas which may still be regarded as ur ban. This combination creates the principle problems of urban recording, and t h e particu lar set of data available to the SMR. This definition would exclude new * West Midland Sites and Monuments Planning & Architecture Department 3, St. James' Road, Dudley DYl lHZ West Midlands

Record

41

N. A. R. LANG

~

~

j !!

II)

I

.. J J i i I ~ ~

~

0

Q.

2 < il

i

l:

i!

1

!

-:--!:

•

i

I• E ¥l

f

--

i

....

~ Q.

'i

l

C

i

;;

,s

0

ll

!i

I

!0

f

!

C

i

:,

{

j

"i

l u 0

i

J

~

,.

0

u

r

..

l

0

! ]

j

u

u

E

j

i:

i

;

•!

'i !

i

~

0

0

i

a:...z

.. .. . ..

'i

. . :.

~

..

i

j !

0

~

~

~

... 0

i


u C"d .... c::

0 N i-.

0

..s::

.... C"d Q)

0. c::

:.a .... ca u ·;:: i-. Q)

>

A = 90 ° 0.033 0.041 0.073 0.005 0.624 0.005 0.073 0.041 0.033 - 0.016 0.035 0.151 - 0.936 2.097 - 0.936 0.151 - 0.035 0 .016

A = 60 ° 0.010 0 .014 0.032 - 0.065 0.586 0.053 0.100 0.060 0.050 - 0.002 - 0.160 0.410 - 1.404 2.009 - 0.561 0.084 - 0.016 0.030

PROFILE CURVES

A = 30° - 0.006 - 0.009 O.Oll - 0.142 0.491 0.088 0.094 0.061 0.053 0.044 - 0.300 0.664 - 1.654 l. 791 - 0.324 0.049 0.003 0.055

B

10

of D-pa rameter

A = 0° - 0 .008 - 0.022 0.007 - 0 .170 0.454 0.089 0.093 0.053 0.049 0.085 0.403 0.808 - 1.751 1.704 - 0.226 - 0.040 0.117 - 0.015

Table 18.1 : Inverse filter coefficients for a magnetized hori z ontal cylinder and verti cal thin plate for different azmutes of profile A. Number of coefficients : 9

171

OLGA KAROUSOVA

AND MILOS KAROUS

--.

---

-.:::: II

H

-!H ++ ~,_~ · ++ ·+++•++++++++++++++++++++++++++++•++++++++++++++++••+•++++ + + •• .,_,................... ,_,_M · +++++++++++++++++•++•++••+++++++++++••+++••++•+•+•••+•+++ +• ._ ++++••••++••+++•+++•+• i-!i- .... :::t••++•+++++++++++++++++++++++++++ •+++++ •••••• +•++++++++ . ..,__. +••+++ ••+++•••++•++•+++•++++++++++++++ i,,,.t,.. .... - +••++•+++++•++•++u++++ u++· +••+•+ ·••+••······tt····••+• • ♦ ••++•+++•+••••++ ++•+ ++ I +•++••+++•++ +++++•++ H--++••++ ++•+++••++ ···••+++++•+••++••++ ♦♦ g++ . + .....~ • • • + +++++• +++• +++++• ++ ++ ++++++•+++ ++++•++++•++•••+••+• + ♦ •• + + '"1..., H+ + ♦ ♦ • +++++♦ ♦ + ♦ ♦ + ♦ + ♦ • • ♦ ♦ ♦ ♦ • +-+-H ++•+•• ♦ +++++••+++++• ._ + + • + • • -+ + + + + + • • •• + ♦ + + ...,........ ++•• ++++++ ++++++ ♦ +••++++• +-+-H +•••++ ._H ~► + +••·•u+•+•+•••tt+••···u•+++ ++++++ •• +•++++••++••++• +• ._ H ~► ++++++ ++++++•. ++++++++ ++++++ ++++ ++++•• •• ++ •• ++ •••• ;..,H t1t++•• ++++++++++++++++++++++++++++-+++•++ ++++++++++++•• ++ ;..,H t1 t+ + •• + + + + + + + + + + + + + + + + + + + + +. + + + + + + - ++++++++++++•• +-++•++ +• I I I +++•++ · + . ...,,__ ++++++++@++++••++++++••I •++• •• •++ . +•++++· +• ++++++ ++ ►----- +++++-+++ ++++ ♦ •++++++ ♦ •I I I I ++++++ ♦ •++ ♦ O ♦ ♦ •/4 + + ++++++ · + ---++ - ++p++++++++++++++•+++ •• ++++++++. ++++ +-+++++ ·+ ++ - ++ ++++++++++++++++++• •++++++++• •++ +++• .++ + ++++•+ ••++++ ++ ••+++•+•+••••+u••+•u···••++• ++++++ -+-+ . + + + +++•++~+++++++ +++• ++++++++ ••++++ ++ ++++++++++ ·· +++++ ++••++••••+++++++•++•••+•+••++ I • + •• + + •• ♦ •++++++++++• ++++-++++++ · ++++• •++++••++ ♦ ♦ + + ♦ • + •• • + + +++ ♦ ++++++ . • + + • .• ++••++• +• . • ♦ ♦ + +. • + + ♦ ♦ + + - ++ ++++++++++ +•+•++• · •++•• . J. . ++++-++++++ · ++++•++ I I••,IT '.I ·1 I I ++•+++++++ · +++++++ ++ 1 1 + +-,-- ++ I IT ·1 I I ++++ ++ + ♦ + -+ t-t + + I I I 1 I I I 1 I I H++++++++ +++++++-+ ++ +•++ ,- ++I I ++++++++++ ++ ++•+ •+++ ++ I I I I ++++-+++++• ++++ .....+ ♦ I I ++ •+++ +• tt +-+++-+ ♦♦♦ ++ +•++++++++ •+++ ti•++-+++++ + ++ ♦ ·- I I ,___ • ++ • +-++-+++++++ I+ ♦ -++++++ ++++••++++ •+++

+-+••······ ·••· +... • • • • . + ••

•• •++++++++++++ ••·••+••

---- ♦ ·••++••

••

••

++++++••++• ~--~---

::j:••··p················••,'''

....... .

~---~ --~ ...... .,

--·+++++++++•+tt•···g······g···· --__ --_.__ ·+-"·. ++++++++++++++++ti•+ -,-® __ -•~

•• •• ...

·••

•••

----

••

ill++++++++ : !!: ! .t-Hi,-

Figure 19.2 : Hatched map

•

•••

••

♦

••

--

ofBA T anomalies

••

•

•••

-

•

••

.t..

+-+

•• •• ~--

..

••

--

--

from Fig. 19 . 1

•• T

T

19.

METHODS FOR PROCESSING DIGITAL GEOPHYSICAL DATA IN ARCHAEOLOGY

Figure 19.3 : Axonometric representation Egypt - 1984 (Eisler et al. 1988)

of Raneferef's 183

mortuary

temple,

Abusir,

V. HASEK AND R. VENCALAK

• computing 1949).

derived

anomalies

according

to weighed

moving averages

(Griffin

The aim of the separation is to obtain from the measured and processed data the geophysical effects of the sought objects. Other components of the field are regarded either as a manifestation of the deeper geological structure, stratigraphic and facial change in the rock massif, or as random measurement errors (Odstrcil 1989). The computation of coefficients of filters for data smoothing and the computation of higher derivatives are based on the approach of G. A. Ljubimov and A. A. Ljubimov (Ljubimov & Ljubimov 1983) . Fig. 19.4 shows the construction of a map of second derivatives from the site of the Raneferef's mortuary temple near Abusir overlaid with a simplified ground plan of the excavated object. The temple elements correlate well. Residual anomalies can be computed as the difference between the measured field and the regional anomaly express the regional field as a weighted mean value in the plane field S, where the weight of the indivdual values depends on their position from the centre of the field in which the computing point P lies. To compute the regional anomaly, we thus (Mares et al. 1979) : !:::.Treg (P)

= -µl -IsI 6.T

(M ·)/J, M

ds

where M is an arbitrary point in the field S, 1-t(M) is the weighting function at the point M for the field S andµ = J J µ(M) ds is the normalisation coefficient. For the numerical calculation we set the weighting function p.(M) = 1, i.e. we ascribe the same weight to each value of the field. The field S is then a circle with radius R = ti\!'5 (where t:::.is the measurement interval). Then the regional field can be expressed as a function of average values on several circles with radius R. Figs.19.5-19.7 show comparisions of radii R used in calculating arithmetic means of anomalies and residual t:::.. T anomalies from the site of the defunct medieval village Srnavka near Lelekovice (Brno district). They demonstrate that the effect of the filter applied in the processing depends on the ratio of the area of plane field S and the area of the anomaly induced by the investigated archaeological object. It is obvious from Fig. 19.6 that if, for example, R = t:::.Js is applied, then in the map of residual ti T anomalies mainly noise is separated. All information about the environment (the useful component) is contained in the map of "regional" anomalies. With larger R, e.g. R = 2 t:::..Js (Fig. 19. 7) the constructed maps, despite various accompanying fictitious anomalies of negative sign, clearly show the local effects of objects of interest on the background of "regional" anomalies whose geological sources presumably occur at greater depths. However, the effects of major archaeological structures such as a complex of buildings or layers of loam with objects may be observed. Some of the programs can be successfully used for processing profile geoelectric measurements, namely for smoothing profile curves, or for cal cu la ting higher deriva tives.

19.4.2

The method of cross-correlation ing

coefficients and controlled filter-

For quantitative interpretation programs were written based on the method of cross correlation coefficients, calculation of anomalous values by the method of ampitude filtering, the method of main components, and controlled filtering. If a local anomaly is clearly oriented in a particular direction, method of cross correlation coefficients can be applied. It is based on stepwise comparison of the

184

19.

METHODS FOR PROCESSING DIGITAL GEOPHYSICAL DATA IN ARCHAEOLOGY

Figure 19.4: Map of second horizontal derivatives from the site of Raneferef's tuary temple near Abuiser compared with the excavated object groundplan

185

mor -

V. HASEK AND R. VENCALAK

1•0

r

·ic JO

+

-

'

'

'

.,

-

10

.··-..

·

" , " f ~ T anomalies. Figure 19.5 : .MaiB~no district) near Lelekov1ce '..,"°

1)0

. o f defunct o n the site

. I village Srnavka med1eva

120

.

L

".

.

c• • . .

l

,·.i :... · ..

..

10 .

'::~ . :_ .) J;

. \\

® •, ..... , 10

--·

.•",

·-._

'••

..:

·_

.

•

,'

;

\

\

'

I

:

\

'

)0

2

(~_! +·.

!. . a of centered Figure 19 .6. M Ir Lelekovice site Srnavka ne

upper and residual lower

186

~T

anomalies

for

•j

R--

)_[_ ~

VS,

-10

19.

METHODS FOR PROCESSING DIGITAL GEOPHYSICAL DATA IN ARCHAEOLOGY

110

1)0

100

'°

,0

10

JO

,0

20,

/\ - -

XI

20

I '

I '

I I

'

10

'

I

:,

:•: 8 ♦

,1 1

•

1 / ,-"'-

'\'--'·~:'\

':

•~-.

'

'

'•

', _ ,.

, ♦',

e

,

~ ~ ------+--'--"~ ~ ~./:......_ .

i .,...L-,l'' .!S:1----"'-"-'-'.~ ------,---'----

110

\

I

·-:_:: o·.~ i .-. 110

~

.,'/\ - /

I

TlO

'

,0

~

t::

UO

I

'

'-.:• ,:\.-

"' :,

I

'

I

''I

10

100

90

to

---.--

~ 70

-~

--.\-·,

) , • 40

Figure 19. 7: Map of centered upper and residual lower 6 T anomalies site Srnavka near Lelekovice

187

:0·'~ 0 --:' JO

for R

JO

= 26

10

y'5,

O

V. HASEK AND R. VENCALAK

effects of inhomogeneity in a particular region with the measured field (Zahara in Hasek et al. 1985). The standard can be a theoretical curve over the sought object or field values measured in places where the position of the sought object was confirmed by another method. The method can also be applied to derived fields. The output is a matrix of cross -correlation coefficients whose columns correspond to individual profiles. The maximum values of correlation coefficients whose position on a profile always relates to the middle of the sought sections indicate the continuation of the observed anomaly on adjacent profiles. The formula for calculating cross -correlation coefficients (R) between the standard field e(x ) given in a region with I( points and by a profile section i with the mid -point J.

where

R-i,J . is the coefficient of cross-correlation

eJ·

e x i,J · . x t,J ·.

between the standard field and the profile section with mid -point j in a window with K point, is the standard field given at points -(I( - 1/2. - 1) / 2 + 1]~ X ... 0 .. . [(I( - 1) / 2 - 1]. ~ x , (K - 1)/2~ x, is the mean value, are values of the measured or derived field on profile i at points xi, - (K-iJ/ 2 ~ x ... xi,j ... x i, (I( - l )/2~ is the average value of the field on profile i and in a window with K points and middle at the point j. ~ X , [ -(I(

x

,

Fig. 19.8 shows the correlation of the Neolithic sacral ring object near Nemcicky by the method (I( = 5). The algorithm of controlled filtering for a complex of two geophysical methods consists of calculating and comparing the correlation coefficients for anomalies and noises within the chosen region with N profiles and m measurements points on each profile (Sikorskij 19 79). The reliability of separating geophysical anomalies depends on estimation of cor relation coefficients of the useful signal. The dimensions of the chosen region are determined by size of anomalies. If the correlation dependence is strong with useful signals and weak with noises, it can be said that the anomalies correlate. The existence or non -existence of an anomaly is determined by an analysis of the parameter

Fig. 19.9 shows an example of the complex evaluation of DEMP and micro -gravity measurements in Zbysov near Rosice (Brno district), where the methods of calculating correlation coefficients and of controlled filtering were applied in a search of cavities due to mining activity. The positions of mining galleries are evident from the ·results.

188

19.

METHODS FOR PROCESSING DIGITAL GEOPHYSICAL DATA IN ARCHAEOLOGY

NEMCICKY KOEFICIENT KORELACE

0

0.

. )7

CJ

~

yr_ ~

~

®

~

Figure 19.8: Map of isolines if cross -correlation district)

189

coefficients,

0~

site Nemcicky (Znojmo

V. HASEK AND R. VENCALAK

A

e

C

0

F

E

G

II

A

e

C

0

E

F

G

A

8

C

E

0

F

A

H

PFlml --

e

C

0

f

G

H

8

A

C

0

H

0

+ u 3 ck) + ·u 23 (.ik) + u 3 , ( ki), "fabr ic-by -form and context 280

26.

ERROR STRUCTURES

=

+

u3(k)

+

Model 5: log mij k = u + U1(i ) + U2(j) + interactions", the saturated model.

Uz(k)

+ U23(jk) +

Model 4: log friijk interactions",

u

+ u 1(i) +

u2(j)

u 23(jk)

+

OF CERAMIC ASSEMBLAGES

u3i .29 .105.46

• 105 ◄ f\C C:C

.105 .54

.105 .72 -700.01 N -200.00E

----25 meters

! -700 .01N -99.99E

PROJECT: TAP'S?

SITE: Non Mak La OVERLAY DATA FILE: NML-COMPASS.txt GRID INTERVAL: 10.00 meters

Figure 28.4 : 1 :500 point plot showing elevation labels generated window covers 1 00m x 100m.

by TMA. This map

The display scale determines the type and density of information that can be presented in an effective manner. For survey data, where measurements are generally 5 to 20 metres apart, a 1 :500 window (Fig. 28.4) allows each point to be labeled with its elevation value, but covers an area of only 100 by 100 metres. A 1: 5000 map window covers 1 km by 1 km, but is too dense to display point labels (Fig. 28.5) . It is useful for showing a general overview of the area, and as a working guide as to what areas have received adequate coverage during survey. COMPASS also supports thematic mapping, which is the portrayal of a data set in its spatial context by using symbols, colours, patterns, contours, and other graphic techniques (Robinson et al. 1984; Bertin 1983). Common archaeological research problems such as activity area analysis, intrasite variation, site catchment analyses, and settlement pattern studies are all ideal applications for thematic mapping. One example of thematic maps are choropleth maps, where data values associated with 301

AN D REW W EISS

Non Mak La Survey -200.00N -500.00E

-200.00N 500 .00E

-1200.00N -500.00E

-1200.00N 500 .00E

250 meters

TMA Release 2.0

PROJECT : TAP'87 SITE : Non Mak La OVERLAY DATA FILE: NML-COMPASS .txt GRID INTERVAL : 100.00 meters

.......Museum Applied Science Center for Archaeology (MASCA) , The University Museum.......... May 18, 1989 ... 6 :09 PM

Figure 28.5 : 1:5000 point plot generated by TMA, which covers The location of the 1 :500 plot in Fig. 1.4 is highlighted.

302

1000m

x 1000m.

28.

THE COMPASS SYSTEM

a spatial area are categorized and represented by a gray scale or pattern (Fig. 1.6). This can be a very effective display technique for distribution maps where data has been collected within quadrats or recorded in bulk within excavation units. Fig. 1. 7 compares the procedures for Computer -Assisted Surveying and Mapping with traditional optical theodolite surveying. The overall process, from surveying to finished maps, is estimated to be on the order of two to three times faster as well as being more accurate and less prone to mistakes.

28.3 28.3.1

Applications Topographic

mapping at Non Pa Wai, Thailand

The system as outlined has been used during the 1986 - 1989 seasons for topographic mapping at metal working and habitation sites in Central Thailand, as part of the Thai land Archaeometallurgy Project, a joint University Museum and Thai Fine Arts Department investigation into the prehistoric social and technological development of metal use in Thailand (Pigott & Natapintu 1988) . The research area is located on the eastern fringes of the Chao Praya River basin (Figs. 1.8 and 1.9). All too often archaeological sites in the tropics are covered in thick vegetation . The total station reduces the need for extensive brush cutting and frequent station changes by requiring only a pencil thin line of site to the prism, rather than three stadia readings. The great range and accuracy of the instrument also allow s the prism to be placed on a very tall pole to get above the vegetation or the instrument to be set up at an optimal station and cover a very wide area . During the 1986 and 1987 seasons the sites of Non Pa Wai, a 5 hectare copper smelting mound, and Non Mak La, a 35 hectare industrial and habitation site were mapped. In 1986 about 30 hectares were mapped around Non Pa Wai over 3 5 m a nhrs, with approximately 1000 points being recorded, while in 1987 45 hectares were mapped at Non Mak La, recording over 1200 points in approximately 50 man -hrs . A 1 :2000 point plot with elevations categorized into contour interval values was made using TMA, and a working map with a contour interval of 50 cm . was produced in the field. This was immediately used to map environmental, geoarchaeological , and cultural data. In Philadelphia the data was replotted and drafted to produce a final map of the entire 75 hectare site complex (Fig. 1. 10).

28.3.2

Systematic surface collection

at Non Mak La, Thailand

While mapping the 35 hectare site of Non Mak La during the 1987 season it was observed that the surface remains occurred in discrete but overlapping distributions of copper and iron slag, crucible and furnace fragments, and domestic earthenwares. In order to investigate these distributions a systematic surface collection was under taken. Transects were laid out by eye and tape to cross the major observed distributions, and 92 collection points were set at 30 metre intervals. These points were then quickly shot in with the Total Station. At each point a circle 1 metre in radius was inscribed in the soil using the dog -leash technique advocated by Binford (1964), the ground cover was carefully removed, and all artifacts were collected, sorted, counted, weighed, and recorded on a field form. Pottery and metallurgical ceramics were then bagged and returned to the field laboratory. In the laboratory three data sets were combined (Fig. 1.11 ): 1. A EDM file of the location of each collection unit, transferred

to EXCEL.

2. The counts and weights of each artifact class, entered manually into EXCELfrom the field recording forms. 303

ANDREW WEISS

Solar lnsolation Values for March 21 Calculated with Program INSOLATN 300 .05N 300.05E

300 .05N -100.00E

Legend

0

>6 .00

□ 5.51 - 6.00 ~ 5.01 - 5.50

[ill

4.51 - 5.00

m4.01 - 4.50 EJ) of the current group. The function J is unity within plus or minus one standard deviation of the mean, otherwise the fuzzy function declines exponentially but is never zero unless the standard deviation is also zero . In formal algorithmic terms this is : LI = m - s L2 = m + s

BEGIN

< /,2 THEN f = I ELSE BEGIN IF s = o THEN J = o ELSE BEGIN IF p