ICL Technical Journal Volume 5 Issue 3
Total Page:16
File Type:pdf, Size:1020Kb
TECHniCAl j o u m n i Volume 5 Issue 3 May 1987 Published by INTERNATIONAL COMPUTERS LIMITED at OXFORD UNIVERSITY PRESS iCL The ICL Technical Journal is published twice a year by TECHniCRl International Computers Limited at Oxford University jouRnfli Press. Editor J. Howlett ICL House, Putney, London SW15 ISW, UK Editorial Board J. Howlett (Editor) F.F. Land H.M. Cropper (F International) (London School of Economics & D.W. Davies, FRS Political Science) G.E. Felton K.H. Macdonald M.D. Godfrey M R. Miller C.H.L. Goodman (British Telecom Research (Standard Telephone Laboratories) Laboratories and Warwick J.M. Pinkerton University) E.C.P. Portman All correspondence and papers to be considered for publication should be addressed to the Editor. The views expressed in the papers are those of the authors and do not necessarily represent ICL policy. 1987 subscription rates: annual subscription £32 UK, £40 rest of world, US $72 N. America; single issues £17 UK, £22 rest of world, US $38 N. America. Orders with remittances should be sent to the Journals Subscriptions Department, Oxford University Press, Walton Street, Oxford 0X2 6DP, UK. This publication is copyright under the Berne Convention and the Interna tional Copyright Convention. All rights reserved. Apart from any copying under the UK Copyright Act 1956, part 1, section 7, whereby a single copy of an article may be supplied, under certain conditions, for the purposes of research or private study, by a library of a class prescribed by the UK Board of Trade Regulations (Statutory Instruments 1957, No. 868), no part of this publication may be reproduced, stored in a retrieval system or transmitted in any form or by any means without the prior permission of the copyright owners. Permission is, however, not required to copy abstracts of papers or articles on condition that a full reference to the source is shown. Multiple copying of the contents of the publication without permission is always illegal. © 1987 International Computers Limited Printed by H Charlesworth & Co Ltd, Huddersfield ISSN 0142-1557 (CL TecHmcfli jouRmi Contents Volume 5 Issue 3 The ICL Fifth Generation Programme Guest Editorial B.W. Oakley 357 Foreword J.M. Watson 359 What is Fifth Generaton?-the scope of the ICL programme B. J. Proctor and C.J. Skelton 360 APPLICATIONS The Alvey DHSS Large Demonstrator Project E.C.P. Portman 371 PARAMEDICL: a computer-aided medical diagnosis system for parallel architectures M.G. Cutcher and M.J. Rigg 376 S39XC-a configurer for Series 39 mainframe systems C. W. Bartlett 385 The application of knowledge based systems to computer capacity management M. Small 404 APPLICATIONS ENVIRONMENT On knowledge bases at ECRC J.-M. Nicolas 421 Logic languages and relational databases: the design and implementa tion of Educe J. Bocca 425 ICL Technical Journal May 1987 I The semantic aspects of MMI J.M. Pratt 451 LANGUAGES Language overview E. Babb 471 PISA-a Persistent Information Space Architecture M.P. Atkinson, R. Morrison and G. Pratten 477 Software development using functional programming languages J. Darlington 492 Dactl: a computational model and compiler target language based on graph reduction J.R.W. Glauert, J.R. Kennaway and M.R. Sleep 509 PARALLEL DECLARATIVE SYSTEMS Designing system software for parallel declarative systems P. Broughton, C.M. Thomson, S.R. Leunig and S. Prior 541 Flagship computational models and machine architecture I. Watson, J. Sargeant, P. Watson and V. Woods 555 Flagship hardware and implementation P. Townsend 575 GRIP: A parallel graph reduction machine S.L Peyton-Jones, C. Clack andJ. Salkild 595 Notes on the authors 600 ii ICL Technical Journal May 1987 Guest Editorial The Japanese announcement of their “Fifth Generation” co-operative project in the autumn of 1981 was the trigger for a rush of other Fifth Generation programmes to emerge in the USA and Europe. In the USA this has partly taken the form of new co-operative ventures, such as the Microelectronics and Computer Corporation (MCC) at the University of Texas, supported by over 20 firms in the industry. This centre is far larger than the Japanese ICOT research centre, established by MITI using staff seconded from the eight large Japanese firms engaged in the programme. And the MCC is by no means the only US Fifth Generation programme. Of course defence money is helping to speed up the work in the USA, notably through the Defence Advanced Research Project Agencies (DARPA) Strategic Computing Pro gramme - not to be confused with Star Wars. In Europe the UK’s Alvey Programme, and the EEC’s ESPRIT programme are the most visible sign of work on the Fifth Generation. But, as in Japan and the USA, all the main computer firms have work going on on aspects of the Fifth Generation, almost by definition of the next generation of computing. And it is often forgotten that the European Computer Industries Joint Research Centre at Munich (ECRC) is a Fifth Generation research centre somewhat larger than ICOT, even if far less visible. This centre is funded by ICL, Bull, and Siemens, and houses a team of high quality. Some of its work is featured in this volume. If it was the Japanese announcement that triggered the Fifth Generation programmes of the Western World, or at least gave publicity to them, it was the recognition that the applications of Artificial Intelligence will lead to large new markets that has led to their funding by firms and governments. Expert Systems are going to become a feature of every considerable walk of life. And the eventual solution of the difficult problems of general purpose speech recognition and Natural Language understanding is going to unleash a whole new range of applications of computers as it becomes easier to “converse” with one’s system. Since the days when Alan Turing was the unsung prophet on the use of computers for symbolic and logical manipulation, not just for the solution of mathematically based algorithms, the UK has had an honourable place in the vanguard of the development of the techniques of Artificial Intelligence computing that lies at the heart of the Fifth Generation. Edinburgh University and other UK research centres have provided a steady stream of ICL Technical Journal May 1987 357 workers in the field, and have pioneered many of the functional and logical language techniques. ICL with its DAP was one of the first computer manufacturers to pioneer the use of parallel computers, which will be essential if the power required to handle all but the most simple of Knowledge Based Systems is to be available. The significance of the DAP is probably more widely recognised abroad, where it has been flattered by being copied, and it probably remains the parallel machine for which the most software has been written. So it is nice to see the leading part that ICL has played in both the Alvey and ESPRIT programmes. This volume brings together articles on a whole series of projects in which the company is engaged. Seen together it makes a most striking collection, both for its breadth and for its quality. The way in which ICL have developed their close links with the excellent work of the British Universities is clear from the list of authors. Is there any other firm in Europe who could put together such a record of their work for the Fifth Generation? It must give the Japanese food for thought! Brian Oakley Director, The Alvey Programme 358 ICL Technical Journal May 1987 Foreword Every industry evolves by a process of steady development - interspersed with major discontinuities. These occur when technology, industry and customer understanding and the economics of business coincide to cause a jump forward to the next plateau of development. The so called Fifth Generation potentially represents such a discontinuity in the Information Technology industry and in this issue of the ICL Technical Journal we examine some facets of this against the background of the overview of the ICL programme in this area as outlined in the first paper. The real and valuable result of any discontinuity will be the impact these technologies have on systems applications. It is in this area that Fifth Generation techniques demonstrate the economic value over alternative systems approaches. The first set of papers describe some current applica tions using these new techniques. To achieve this, some radical rethinking will be required in the applications environment - the way the system and the user interact. The set of papers from the European Computer Industries Research Centre (ECRC) describe the basic research goals being explored in this area. The key technology streams of new language types and the development of more powerful and parallel machine architectures provide the underpinning technical capability and the papers in this area draw heavily on the work carried out under the UK co-operative research programme - Alvey. We are making no predictions as to when Fifth Generation technology in total will have major user impact, however what is clear is that the ingredients for a substantial step forward in the quality and usability of information systems are now coming together. J.M. Watson Director, ICL Marketing and Technical Strategy ICL Technical Journal May 1987 359 What is Fifth Generation? - the scope of the ICL programme B.J. Procter and C.J. Skelton Mainframe Systems Division, ICL, West Gorton, Manchester Abstract The scope, motivation and modus operandi of the ICL Fifth Genera tion programme is outlined to provide a context for the remaining papers in this issue of the ICL Journal. 1 Introduction The single chip microprocessor has brought a computer-based revolution to the 1980s. In human terms, however, the applications we run on these processors are extremely simple.