Jump Over Left Menu
Chapter V: Meeting the Needs of the Universities and Similar Institutions
148. In this chapter we make recommendations which, we believe, would substantially meet the needs of the Universities and Colleges of Advanced Technology by the end of the five year period. However, the pace of development of computer technology is such that recommendations for all but the first two or three years should be regarded as tentative, to be adjusted according to new information, both of the requirements and of the equipment, as it becomes available. In order to envisage the scale of a suitable programme we have had to suggest financial estimates, however rough, to every relevant institution; we have had to base our estimates upon known equipment and at present prices. Our recommendations are thus firm only for the first two years and for one or two key matters thereafter; for the rest, it is clear that our findings should be regarded in the spirit of a first attempt at a rolling five year forecast. At the end of our Report we make suggestions regarding the continuation of our work in future years.
149. It is clear that it would not be economic to provide every University with the largest available computer. It is equally clear that any research worker may occasionally need to use a much larger machine than has been provided locally. Indeed, one of the reasons for setting up the Chilton Atlas of the Science Research Council (formerly of the National Institute for Research in Nuclear Science) was to make available on a national basis one of the largest machines obtainable at that time. Although several years of useful life can be expected from this expensive installation it is already unable to satisfy the country's need for large-scale computations.
150. We have therefore given particular attention to the problem of how to make available a number of large installations to all research workers in Universities and Research Council laboratories. We have come to the conclusion that a number of regional hierarchies would most readily serve our purpose. By a regional hierarchy we have in mind a group of nearby Universities and research laboratories, most of which will be provided with modest machines of their own, sharing a much larger facility provided at one of them. Bearing in mind the rate at which computing demand is known to increase it might often be feasible to allow the regional machine at one location later to become the local machine, being replaced elsewhere in the region. We propose three regional machines for immediate consideration at London, Manchester and Edinburgh, which might be identified with the South-East, the North-West and the Scottish regions respectively, but we consider that attention should eventually be given to further regional machines in the South- West, the Midlands and the North-East. We do not believe that there is anything particularly revolutionary in this proposal for we have found much evidence for informal arrangements of this nature already in operation all over the country, and we merely wish to strengthen and speed the process of natural growth in a rational manner.
151. The regional machines should not only provide a fast turn round service in the local University but facilities for adjacent Universities and for nearby Research Council activity. The regional centres should be staffed so that expert computer knowledge, particularly in the industrial application field, can be made available rapidly and effectively.
152. It should be added, however, that we would generally wish to discourage an alternative arrangement of a less healthy variety. It is sometimes. proposed that a Research Council should possess a computer hierarchy of its own in spite of the fact that its laboratories may be spread widely over the whole country. We see no point in such an arrangement: indeed, we believe it would foster a narrow outlook and hinder the proper development of computer techniques through lack of essential contacts with other disciplines; moreover, inevitable difficulties of communication and travel would, we believe, cause such an arrangement to break down in favour of informal local arrangements.
153. It should also be stressed that the provision of a large regional computer should not be made an excuse for failing to provide local computing facilities adequate for most normal needs, for the large computers should not be over-burdened with small jobs. The hierarchical principle can be carried a step further by the provision of smaller departmental computers within a University or a Research Council laboratory where there is clear evidence of sufficient demand which would otherwise over-burden the central facility. Financial stringency has made it impossible for the U.G.C. to make grants for departmental computers in the past, although a few have been made available by Research Councils for special, usually on-line, applications. We recommend that this policy should now be relaxed, although we make no specific recommendations for departmental machines in this Report.
154. Acceptance of the hierarchical principle brings to the fore the problems of compatibility of computers and of programs, a matter with which we have already dealt in detail. We recommend that every computer provided under this programme should wherever possible be capable of operating at least on the two major languages Algol and Fortran; where a system has magnetic tape units one at least should be IBM compatible; and they should all be provided with card readers and punches as well as paper tape facilities.
155. This will, at minimum cost, ensure a flexible and efficient arrangement by which any computer may be used as an off line input-output device to any other, programs and data being transmitted (by car or otherwise) on standard magnetic tape. A group requiring extra computing power can then obtain it quickly and without hindrance, and expensive installations can be used to the full. Programs and data (e.g. satellite tapes) can also be exchanged and distributed freely, and collaboration with other countries facilitated. Special equipment such as graph plotters, microfilm records and readers which may be possessed by some groups can readily be made available to others. Particular attention should be paid, in discussions with the manufacturers and in setting technical and contract specifications, to ensuring that true compatibility with international standards is achieved, since difficulties are frequently encountered in this respect. Consideration might be given to making this the responsibility of the Computer Advisory Service of the Ministry of Technology. On a rather longer timescale, consideration should be given to international standards for magnetic tapes, and other forms of mass storage and to automatic data transmission.
156. A large regional computer, with power perhaps fifty times that of the IBM 7090, is capable of servicing an equally large number of line printers, and to a lesser extent, of card readers. These should not all be situated at the machine itself, because of the difficulty of sorting and transporting bulky material such as paper and cards. As a general rule, such material should be dealt with as close to the actual customer as possible by means of satellite equipment, and the information transmitted to and from the main machine via automatic links or IBM compatible magnetic tape.
157. In considering how computing needs should be met considerable attention has been paid to two aspects of modern automatic computation which we consider vital. These are:
- The foreseeable massive increase in demand for computer use in medicine, the social sciences and in the non-physical sciences generally, which brings clearly to the forefront the punched card system rather than punched paper-tape system.
- When an installation becomes overloaded there should be a bigger system to which the installation can be upgraded so that good compatibility and continuity can be maintained.
158. At each University we have considered not only the student population and the location relative to other Universities; much attention has been paid to the background history of automatic computation at the University and to the type of work for which extra facilities are needed. Our experience is that efficient service is achieved at rather few installations and that generally it has to be backed by a long history of computer development. We also took into account the known usage of the national Atlas facility at Chilton and of other large machines external to the University in assessing both need and experience.
159. There is one special difficulty that arises in Universities which have made substantial use of Atlas Autocode on the Manchester, London and Chilton machines and which may therefore have large numbers of programs already written in this language. It is evident that none of the new series are to be provided with compilers for this language and it will be necessary for those Universities wishing to continue its use to provide a compiler on their own or in collaboration with other similarly situated institutions. The Working Group sought out the views of the computer manufacturers on the provision of compilers for institutions whose programmes are in Atlas Autocode. It appeared that if Universities wished to continue to use this language they should consider producing their own compiler.
160. Special attention has been given to the needs at the proposed regional centres. It has also been considered axiomatic that there should be a substantial computer installation within each of the other broadly recognisable regions, and that moreover a regional machine should be made available to as wide a usage as possible within the Universities and Research Councils. Universities situated remotely have received special attention and in some cases a bigger installation has been suggested in order to combat the severe lack of access to regional centres. For Universities reasonably close to each other but at great distance from regional centres attention has been paid to local compatibility of the proposed machines.
161. Before turning to the provisions we recommend should be made at each institution we wish to draw attention to a matter that we consider to be of great practical importance. We found that there is at present considerable duplication of effort amongst the Universities and the other research institutions; in many cases we discovered that similar programs were being developed at several places simultaneously, to solve much the same problem on much the same facility. One example is the matrix manipulation associated with stress analysis in structural engineering. Had it not been for intervention by the Chilton laboratory another would have been the analysis of X-ray data in crystallography; duplication in the development of film analysis programs for nuclear physics has also been avoided thanks to the coordinating activities of the Rutherford Laboratory.
162. Although some duplication is unavoidable and even desirable there is no doubt that too much manpower and machine time is being needlessly wasted in this way. We recommend that the Research Councils should take a definite lead in drawing together the country wide efforts on the development of major programs of a standard nature. Much can also be done within each University, for example, by maintaining an up-to-date library of standard programs and by collaborating with other Universities in doing so.
163. It was clearly evident from our talks with the computer industry that there were rather few things that could be done quickly, that is within 12 to 18 months. The English Electric KDF 9 installations could be upgraded in this time and a few other machines of British manufacture could be supplied. The delivery times on all the new series are generally more than 18 months, especially if a large number are ordered. These series have mainly been suggested for institutions needing to install moderate facilities after the middle of 1966, or to replace out of date equipment. Modular equipment of this kind has the great advantage of being capable of large-scale progressive expansion, a feature of which we have taken advantage in spreading out the provisions in some cases.
164. It is also clear from our discussions that the very large computers required for the proposed regional centres at London and Manchester must be of United States manufacture if only for reasons of size, at least for the time being. The choice is between the CDC 6800 and the IBM 360/92.
165. The main body of small and medium sized computers required by the Universities can, and we believe should be supplied by British industry except in the few cases where American equipment has special advantages which make its installation substantially more suitable or a more economic proposition. In order to meet the programme quickly we have attempted to spread the load evenly over the various manufacturers.
166. It is considered to be essential that the provision of public funds should be adequate to meet the computing needs of the Universities. The Working Group considers it inadvisable for Universities to install computers as a quasi-commercial undertaking or to attempt to pay for part of the cost of a computer by selling time commercially.
167. Our recommendations are summarised below; the notes which now follow are intended to explain as far as is possible in a few lines how we arrived at each suggested installation. We repeat that our suggestions are tentative for the last three years and in general are intended to set the scale of overall expenditure. Where specific computers are named, this is by way of indicating the general kind of computer intended, and is, for example, without prejudice to proper tendering action.
KDF 9 Universities
168. Top priority in time must go to upgrading the KDF 9 machines already installed. English Electric have shown that they have sufficient production capacity to support this computer which, after a distinctly uncertain beginning, has now settled down to become in effect the backbone of computing in British Universities. The Company have convinced us that they are able to deliver within 18 months sufficient equipment to upgrade all these machines up to Culham standard, i.e. 32K store together with adequate tape decks, disc storage, and peripherals. The advantages of upgrading are clear and simple: the frustrations associated with inadequate core storage, too few tape decks and adequate peripheral machinery will be met; moreover the upgraded machines will accept the two main languages Algol and Fortran. Coupled with this improvement in hardware, however, we make a definite recommendation that the contracting company should produce an efficient Algol compiler for the upgraded machine.
169. The Universities concerned are Birmingham, Glasgow, Leeds, Liverpool, Oxford, and Salford R.C.A.T. The Nottingham machine, not yet delivered, should also be upgraded, if possible before delivery. We understand that each of these seven institutions will be prepared to accept our recommendations, although some of them (Liverpool, for example) will need financial assistance with building modifications that will have to be put in hand immediately, and with increased staff. The remaining University with a KDF 9 machine is Newcastle; their building programme is such that upgrading could not be performed under 2t years. We accept their view that it would be better to install an entirely new machine of a different kind after that time. The upgrading of the Salford R.C.A.T. machine must be considered as part of the plan for meeting the immediate needs of the Manchester regional hierarchy.
170. The total cost of the above proposals excluding costs already approved has been spread over the first two years, those institutions coming in the first year having to our minds some additional degree of urgency. We recommend, however, that the operation in order to be conceived and executed as a whole should be coordinated in collaboration with the English Electric Company. We understand from the Ministry of Technology that it is desired to upgrade the KDF 9 machine; at the National Physical Laboratory as part of this programme (although its cost falls outside it), and in the next chapter we recommend a new upgraded machine for the Medical Research Council.
171. Each upgraded machine provides extra computing power comparable to that of a 7090 for an additional cost only a fraction of the 7090 cost, and there is no other British or foreign computer at present available of which the same can be said. Furthermore, a body of customers is created, extending across Universities, Government Departments, and the U.K.A.E.A., large enough to stimulate and encourage further hardware and software development by the Company.
172. At some stage the KDF 9 Universities will require replacement machines. It is our present view that Oxford deserves some priority in this respect, and in token of that we have included a sum for Oxford in the year 1969-70.
Proposed Regional Centres
173. Existing facilities, principally the Atlas computer of the University of London Institute of Computer Science and the IBM 7090 at Imperial College, are already heavily loaded. A good deal of London University computing work is done on a variety of other machines.
174. There is an urgent need for the provision of further facilities. A number of constituent colleges have convinced us that their essential needs cannot be met unless they possess machines of their own in addition to central facilities, in order to have the advantages of immediate access, quick turnround, and experience in machine operation. The needs of Imperial College, University College and Queen Mary College seem especially pressing. Relations between the various colleges, and between them and the University itself, are not a matter in which we would wish to intrude too far. Suffice it to say that we have given the matter very close attention and that we suggest, subject to the agreement of the University and its constituent parts, and within the total estimated funds we here recommend, that the best solution might be as follows.
175. It is proposed, in order to obtain rapid relief for London, that a sum should be allocated for the provision of an IBM machine to be installed in 1966 at University College. This machine should be fully compatible with the regional machine proposed below, and a substantial fraction of its time should be made available to other University users. Early installation of this machine, together with the release of further Atlas time, could bring about a substantial amelioration in the immediate computing situation in the London region. It is further recommended that sums should be allocated to Queen Mary College for a three-year development beginning in 1967 (we have notionally assigned this sum to the financial year 1969-70) and that additional sums should be made for the provision of machines to other constituent colleges in consultation with the University as a whole.
176. In the longer term the problem can be solved only if London develops rapidly as a regional centre serving all institutions in that area. We therefore recommend that a substantial sum should be provided for a large machine in 1967-68. It is suggested that this machine, for which the only contenders at present are CDC 6800 and IBM 360/92, should be sited at Imperial College.
177. The University has taken a leading part in world development of large digital machines and thereby has provided its own facilities since 1951. The Atlas is performing satisfactorily, although hampered by shortage of space and money; present use runs at well over 100 hours per week and the difficulties of sharing the machine with ICT are recognised. It does not seem likely that there will be any substantial improvement to the performance of the machine, but the University's application for disc file storage would certainly help to smooth work through the machine, so increasing its efficiency, and a sum is proposed to meet this need immediately.
178. The present Atlas contract with ICT comes to an end in 1970 and a new building will not be available until 1969. We welcome the present cooperation between the Faculty of Science and the Faculty of Technology (The Manchester College of Science and Technology) and their willingness to provide a wider service. It is therefore proposed that a large American machine be installed in Manchester on a hired basis as soon as it can be accommodated, that is in 1968-69, and that this be operated as a regional centre. We suggest a provisional sum for the years 1968-69 and 1969-70 for this purpose. The regional machine should provide a general service for the University as a whole, and also for the Manchester Teaching Hospital, other Universities in the North-West as required and as of right (e.g. Bangor, Keele, Lancaster, Leeds, Liverpool) and nearby Research Council activities.
179. The need for a major installation at Edinburgh has already been acknowledged by Sir Willis Jackson's Panel. A major decision of the Working Group is that Edinburgh, because of the nature and volume of its own demands :and also because of those of the many nearby Research Council laboratories, is .an excellent case for treatment as a regional centre along multiple-access lines including the provision of one or two small satellite computers. We also propose that there should be some coordination between Edinburgh, Aberdeen .and St. Andrews as part of the regional nature of the Edinburgh installation. The Working Group considers that Edinburgh could take a leading role in the development of conversational systems as a matter of national importance. Much of their task would initially be concerned with the administration and planning of what should be an advanced computing network for the considerable and expanding Research Council activities in their neighbourhood. Their role will be a difficult one, requiring tactful guidance and coordination as well as technical skill. We have noted one company's interest in this tentative proposition, and their desire to undertake it. We have no evidence that any other company would be able to satisfy an adequate time-scale.
180. We therefore recommend that discussions should start as soon as possible between the University, the Research Councils, and the company to determine how a conversational system could best be implemented in detail, :and at what precise cost. We have provisionally suggested sums for 1966-67 :and for 1969-70 additional to any previously considered sums.
181. The University has pioneered large scale computation and computer technology for many years. The Working Group feel that a new project at Cambridge, for which the University has expressed a need without estimate of size or cost, should have their support, and propose that a substantial sum should be made available to the University during the period 1969-70. It is yet too early to define how this money should be spent. In the meantime, we recommend that sums should be provided in each of the years 1965-66 and 1967-68 for development of the existing machine, including the addition of a further 32K of core store.
182. The Working Group would like to see a major installation in the University and to see two links established for the benefit of Exeter and the Bristol College of Advanced Technology. To this end an early extension of the Elliott 503 is proposed for 1966-67.
183. The University should begin to consider in the light of their experience in the past two years and of their position in relation to the South-West, the acquisition of some new computer system for installation later on. A tentative estimate is suggested for 1969-70.
184. The Working Group agree that the Mercury, while being run well, is very fully loaded and inadequate for University needs. We recognise that their application to the U.G.C. has been held up pending our Report. The ICT 1907 for which the University has made application cannot be delivered before December, 1966, and that assumes a perfect scheduling at I.C.T. We point out that the configuration requested is unsuitable for medical statistics which is one of the main points in their argument for improved facilities.
185. However, we agree that a suitable ICT 1907 configuration is the sort of computer that should be installed and a sum is proposed for 1966-67 in excess of their present U.G.C. application; the provision of card input/output equipment should be incorporated at once.
186. The Working Group is entirely in sympathy with the need for a modern computer at Belfast. The present facilities in Northern Ireland generally are very weak and the use of computers in England and the U.S.A. by Belfast has been carried on under great difficulties with much zeal and considerable success. Their use of the Chilton Atlas, for example, alone, shows clearly that the computation requirements at Queen's University are substantial. It is recommended that their application for ICT machinery be approved. The Working Group is anxious to see a computer established in Belfast quickly and therefore recommend that an ICT 1905 configuration should be installed as soon as possible. Plans for further improvement of computing facilities in the University should include two further upgradings to a sizeable 1907 machine by the end of the five year period. The financing of this would fall within the responsibilities of the Northern Ireland Government.
187. The upgrading of the KDF 9 computer which would have been recommended to bring the University machine in line with the other KDF 9's is not practicable because the University cannot implement it for more than two years. We therefore recommend that the University should plan the development of a new computer centre and should investigate the possibility of providing a regional computer service for Durham and for possible future developments at Sunderland; sums are proposed for this in 1967-68.
188. We acknowledge the urgent need for a computer at Aberdeen to provide capacity for work in the immediate future which cannot be handled on the Elliott 803. We feel that a close link between Aberdeen and the new regional facility at Edinburgh is most desirable, especially for the matrix work in the structural engineering field.
189. The statistical work at Aberdeen together with the projected computer work on medical records does however demand much improved local facilities, and in view of the fact that the University's application to the U.G.C. Panel has already been deferred, coupled with the possibility of early delivery and of a link with Edinburgh, we recommend that discussions with Edinburgh on the Edinburgh and Aberdeen specifications, with an eye to maximum compatibility, should begin at once. We have made allowance for the estimated cost of such a facility in 1965-66.
190. We feel that the considerable expansion of the KDF 9 at Glasgow University strengthens the computing future for Strathclyde also, and they are encouraged to use the Glasgow machine more. We feel that the distance between the two Universities should not be a serious drawback to their own computer development. It is appreciated that they wish to develop a computer centre of their own but we feel that the installation they have proposed is too ambitious at this stage and that a start should be made by installing a less extensive ICT 1905 system. The installation should be planned with an eye to upgrading, and sums are proposed for 1965-66 and 1968-69. The initial use of the Glasgow KDF 9 should continue throughout the period of their own computer build up.
191. We agree that the Stantec Zebra should be replaced and note that the College is to submit a case to the U.G.C. It is felt that a new computer should be acquired quickly and we suggest that a new centre should be developed, but not administered by the Physics Department as at present. The new centre, should be a separate entity set up to provide a general service to the College. It is felt that an Elliott 503 computer should meet the needs for the next few years and an appropriate sum is proposed for 1965-66.
192. The growth of computation at Cardiff should be examined about two years after installation of the new machine and a decision should then be taken on what sort of equipment should thereafter be provided. It is proposed provisionally that a new machine should be made available during 1969-70.
193. The immediate needs at Hull can be met by the present extensions already granted by the U.G.C. to their Elliott 803 machine. The need for more ambitious work in Algol should be considered along with the upgrading of the KDF 9 computer at Leeds. It is proposed that Hull should have good access to the Leeds machine by data link until they can establish a more advanced computer facility of their own. We feel that their future needs are probably best provided for in a machine of the new Elliott 4100 series, although the quicker provision of a 503 system might be preferred by the University. An estimated sum for the year 1967-68 has been proposed for this purpose.
194. The University's immediate requirements have been met by the U.G.C.'s approval of an ICT 1907 system. A reconsideration of their requirements should be made in not more than two years time, when it should be clear what additions or modifications can or should be made. We have noted particularly the special importance of the installation of the granted machine, being the largest ICT 1900 system so far recommended for a British University. There will be much interest in the arrival, installation and operation of this machine; we feel that subject to satisfactory performance of the new computer and reasonable development of the University's computing requirements, additions and extensions might be made to develop their computer during the period 1968-69.
195. The College's claim that the expansion of research activity of computer conscious departments can be expected to be considerable in the immediate with a new computer is considered vital for the College, and the proposed appointment of three assistants during 1965 to supplement the present staff of two programmers and two operators seems totally inadequate for the sort of installation they envisage. Nor are we satisfied with the status of the senior staff associated with the present service. The Working Group note that Swansea have static and dynamic structural problems involving the solution of large systems of linear equations, in common with many other Universities, and we feel that this work can only be tackled properly on a large regional facility or on the Chilton Atlas. It is proposed that the IBM 1620 system be retained for teaching purposes and for on-line experiments. It is proposed that a start should be made to build up a considerable ICT 1900 computer system and to this end two sums are proposed in 1965-66, and in 1968-69.
196. While the University seem to have no great plans for computing we feel that the Elliott 803 system cannot be regarded as a suitable permanent computer for the five year period under review. A survey of their requirements should be made in the light of a new Elliott 4100 computer for installation during the period 1967-68. The University should report their findings to the U.G.C. as quickly as possible.
197. The University are encouraged to review their computer requirements for the next five years. We recommend that they should do so with an ICT 1900 system in mind for installation during the period 1967-69.
198. The present Elliott 803 seems to satisfy the needs of the University for some time. In a year or two however the University should consider making a case for a new computer based on the ICT 1900 series.
199. The Working Group agrees that the University has a serious computer problem and that something should be done to meet their needs quickly. It is also agreed that their considerable use of the Elliott 803 could be greatly assisted on compatibility grounds by the acquisition of a further Elliott 503 system. This is recommended. While it is much too early at this stage to decide on further developments at Leicester, the Elliott 503 machine should be installed with an eye to a further improvement in the central computer facility and we propose a sum for 1969-70.
200. The Working Group notes that late in 1964 the College found difficulty in providing a detailed specification of the computer they would need but indicated that an IBM 360/50 was the sort of machine that might be useful, and wished to have some idea of the amount of money which could be made available. We note the construction of a new building to house the Computer Unit, and suggest that the system obtained by the College should be a new upgradable one. This rules out the Elliott 503 system which has also been mentioned. We feel that it would be difficult for all the constituent colleges of the University of Wales to use one central computer facility and recommend that a start should be made at Aberystwyth by the acquisition of an Elliott 4100 system as quickly as possible. Plans for upgrading this machine should be kept well in mind from the beginning and an upgrading is proposed for 1967-68. Close co-operation between this College and Bangor is strongly recommended.
201. We agree that in two years time the College's computing facility will not be able to cope with demand. We feel that it would be difficult for all constituent colleges of the University of Wales to use one central computer facility. The college is expected to report on their five year requirements to the U.G.C. shortly. The Working Group consider that the College's needs would be best served by the installation of a new upgradable series and suggest that there are advantages for them in the Elliott 4100 series. It is therefore proposed that sums be made available for 1966-67. This computer should be installed with a view to considerable upgrading in 1968-69. The College's computing plans have much in common with those of Aberystwyth and much can be gained from joint consultation between themselves and between them and the manufacturers if our recommendations are accepted.
202. The Working Group believes that while the University's computer needs will certainly grow rapidly, neither the size of the University planned for 1967 nor the nature of the work warrants the acquisition of the very large computer system which they have requested. It is felt, however, that the University should not be left long without some computer facility and we suggest that a start should be made by installing an ICT 1905 computer in 1966-67 with the possibility of some upgrading towards the end of the five year period.
203. We consider that upgrading will be justified by the University but the extent to which it should be carried out cannot be foreseen at this stage. In the meantime they should work in collaboration with Herstmonceux who will have a similar machine, and should make use of the London regional hierarchy.
204. The University computer belongs to Prof. McSweeney as his private property. The Working Group notes with interest that the University has set up a Computer Committee in order to discuss the possible ways in which they might develop a course on computation; it is suggested that Keele should develop a new computer centre by the acquisition as soon as possible of a small upgradable computer. The new computer should be installed so that upgrading can occur towards the end of the five year period; we suggest initially an ICT 1903. It is assumed that the University will have access to Manchester.
205. The Working Group feels that the suggestion conveyed to them that an Elliott 503 machine might be suitable for early installation is not a good one. There is little evidence at present that there will be a heavy computing load in the near future but we acknowledge that a substantial load will eventually be built up at York and recommend the acquisition of an upgradable Elliott 4100 series starting in 1965-66.
206. The University should report to the U.G.C. on the state of their computation needs when this computer is commissioned, with a view to later upgrading. In the meantime we suggest that they work in collaboration with Leeds and Manchester.
207. The Working Group notes that the University is currently formulating ideas on their computer needs and what form their installation should take. The Working Group considers that it will be some time before a grant for the establishment of a computer installation could be justified and suggest that the University should plan to install an I.C.T. machine during 1968-69.
208. The U.G.C.'s recent approval of an ICT 1909 configuration covers the needs of the University for some time to come. The Working Group notes that the University is working on a plan for the development of computer services in future years, and suggest that their facility can be upgraded towards the end of the five year period. They should consider themselves within the Manchester regional hierarchy.
209. The Working Group notes that this University has already received a grant for a computer, and feels that this installation should be sufficient to meet their needs until 1968-69 when considerable upgrading may be justified.
210. It is the opinion of the Working Group that Warwick should build up a computer installation in stages. They should base their installation on a small Elliott 4100 machine and plan to have considerable upgrading during 1968-69. In the meantime they should have access to the upgraded KDF 9 at Birmingham.
211. We note that the University is formulating a case for submission to the U.G.C. for a computer system. As for Warwick, we suggest that the university should plan to begin with a small upgradable system, the ICT 1903 for example, with expansion in mind for 1968-69. They should have access to the London hierarchy.
Queen's College - Dundee
212. The College has a Stantec Zebra purchased from College funds and private donations in 1961. The machine is in use 70 to 90 hours per week and is proving too slow. It is not possible to increase its storage capacity. They feel the need of an Algol compiler; this is not practicable using the present installation. We recommend an ICT 1905 in 1966-67.
Former Colleges of Advanced Technology (other than Salford)
Northampton C.A.T. (London)
213. The Working Group have much interest in the recent ICT 1905 installation. It would be helpful if our successor body received progress reports on the College's experience with their computer system. Subject to satisfactory performance of this early 1900 series delivery we consider that an upgrading should occur in the period 1968-69, and that they be considered part of the London hierarchy.
214. This College should share in the upgraded KDF 9 facility at Birmingham University. They should, however, plan a separate computer facility apart from the recently awarded 803. We believe that a provision for 1967-68 is justified in their case, and would suggest an Elliott 4100.
215. The U.G.C.'s grant providing an Elliott 503 machine has done much to solve immediate computing problems. We acknowledge that a line printer for which they have expressed need is a necessary addition to their present equipment. The College's move to Guildford provides them with an opportunity to set up a new computer installation and we suggest than an Elliott 4100 machine should be considered at that time. In the meantime they should use other London machines and the Chilton Atlas.
216. The Working Group has given the Loughborough application a great deal of searching inquiry. The College's application to the U.G.C. is well supported by evidence that they could make use of a sizeable installation and we have noted their view that the expansion of their computer work over and above the capacity of their IBM 1620 might be best met by American equipment. We also note that substantial administrative work is planned for the new computer. However, we believe their essential needs can be met by an ICT 1905, to be installed as soon as possible.
217. The recent U.G.C. grant enabling this College of Advanced Technology to purchase ICT equipment will meet their needs for some time. University status with its consequent expansion suggests that some upgrading of their ICT system must occur towards the end of the five year period. It is not easy to define the nature and scope of this upgrading; tentative provision is made for 1969-70.
218. The College's computer needs are being met, apart from the need for more peripheral equipment. The College should think of replacing their current machine in a few years time, perhaps a new Elliott 4100 machine would meet their needs for 1968 onwards. They should be encouraged to use the facilities at the adjacent University College in Cardiff.
219. The College's new 803 computer should be working this year and no immediate difficulties are foreseen. By 1967 the Working Group feels that the demand will be well in excess of their present capacity and propose that a new Elliott system should be discussed within the College. They should be regarded as part of the London region.
220. For the present the College claims that analogue equipment satisfies their needs. They appear to have given no attention whatever to digital equipment, additional to their present Elliott 803. Use of the extended facilities at the University of Bristol should be possible, but a digital machine should be made available for the College towards the end of the five year period upon their impending move to Bath.
221. The Elliott 803 system satisfies immediate needs. A replacement should be considered in about one year's time and a new Elliott system is suggested.
222. We have not looked in general at the Scottish Central Institutions, but Heriot-Watt College is about to become a University.
223. The Working Group appreciates that the College's Sirius machine may be heavily loaded. While the College should be able to use the proposed Edinburgh regional facility they feel that a modest new computer should be installed in the College. It is proposed therefore that an Elliott 4100 computer should be studied by the College for possible installation during 1967-68.
The Teaching Hospitals
224. It would appear that much exploratory work needs to be done by the teaching hospitals to define the configurations that would best suit their requirements. The nature of their work is very different and a configuration at one hospital may well be quite unsuitable at another. Two things are clear to the Working Group:
- It is right and proper for a teaching hospital when a computing load can be foreseen and justified, that a data processing computer be installed in the hospital to supply a general service within it.
- Wherever possible the local computer must be compatible either by hardware or software features, and preferably by both, to a larger computer at the associated University.
225. Of the twenty five or so teaching hospitals in Britain it is estimated at this stage that about ten can and should be allowed to acquire miniature or small computer equipment during the next five years. We are of the opinion that a sum of £0.05m should cover the cost of a typical teaching hospital system. Consideration should be given to the provision of about two such machines annually over the next five years. The additional research requirements of teaching hospitals should be considered sympathetically.
The Needs of Technical Colleges
226. The Working Group briefly considered the general needs of Technical Colleges and of the remaining Scottish Central Institutions. Within the time available they were not able to deal with these needs in the detailed manner with which they had dealt with those of Universities, but some remarks will not be out of place here.
227. Technical Colleges in England and Wales are financed by Local Education Authorities but the Department of Education and Science is able to survey and influence the computer position through a regulation which requires that no item of any equipment for Technical Colleges costing more than £500 may be purchased by Local Education Authorities without Departmental approval. At present there are 25 Technical Colleges with computers, the cost to public funds to date being approximately £½m. The Department of Education and Science anticipates that the rate of computer expenditure in Technical Colleges will grow, and that over the next five years the cost will be something like £1.2m. The Working Group has not attempted to assess this figure or to suggest whether it should be altered in any way; it notes that the future needs of Scottish Central Institutions will be additional to any English and Welsh requirements and it is also possible that, in the course of time, a demand will grow for small computers at schools. This we have not attempted to estimate.
228. There are, however, some general comments that can be made about computers in Technical Colleges. The computers used should be compatible with the larger computers which we have recommended for Universities and Research Establishments. There are two reasons for this. Firstly, that persons trained at Technical Colleges should be able, at the end of their training, to use these computers which they will be using in the course of their employment. Secondly, there is the argument that the small computers which will be available to Technical Colleges might not by themselves be able to deal with all of the research problems which could arise there and they should therefore be compatible as far as possible with larger nearby facilities.
229. In addition, such computers as are provided should be able to handle the modern programming techniques as detailed in our descriptions of University computers. We are advised that small computers at present being developed by British manufacturers will be able to do this. Unless attention is given to this point we consider that the educational value of very small computers might be negligible relative to the expenditure involved.
230. The Working Group recommends that the relevant branches of the Department of Education and Science, and the Scottish Education Department be informed of their views.
The Use of Second Hand Machines
231. It is possible that some savings could be achieved by transferring older machines where appropriate to meet the needs elsewhere. In general, however, the saving would not be great when allowance is made for costs of re-installation, and in most cases it would be better to allow such machines to be operated by a department or faculty of the University having special needs. In this way a given computer can be used for its whole working life with a minimum of disturbance. Certainly in the case of a large and expensive installation like the London Atlas, there should be no question of moving it and when it is no longer profitable to continue operation it should be scrapped.
Summary of Recommended Expenditure in Universities etc.
232. The total expenditure on computer hardware, at universities and bodies about to become Universities may be summarised as follows:
Of this sum £5.73 m represents American equipment.
233. In the financial year 1965-6, a further sum of approximately £84,000 will have to be found to pay for two items already approved by the University Grants Committee at Leeds and Edinburgh Universities.
234. These figures assume that all machines are delivered in the year quoted, with no slippage, and are purchased outright rather than hired. There is no allowance made for extra peripherals (normally running at 5 per cent per annum of basic cost), for buildings, maintenance, running costs, staff, etc. All these points are considered in Chapter VIII in detail.
235. In addition, the hardware needs of the teaching hospitals may be estimated at £0.5m over the period, while the total likely level of expenditure on technical colleges and Scottish Central Institutions will be of the order of £1 t m. Again these figures do not include the extras mentioned above.
236. Out of this total, a sum of £0.43m over the five year period represents the cost of a computer recommended for the University of Belfast, which would be financed through the Government of Northern Ireland.