Jump Over Left Menu
Chapter VI Needs of the Research Councils
237. As described in Chapter I, we received oral and written evidence from the four Research Councils regarding their needs, but for reasons of time we have not been able to acquaint ourselves as thoroughly as we would have wished with the special problems of many of the various laboratories coming under their jurisdiction. It seems clear, however, that until very recently insufficient attention has been paid to the need for computers by the Research Councils generally, and this has made it difficult for us to do more than to estimate in a rough way many of the needs. The exception is the Science Research Council for which a detailed picture exists at least for the larger installations. We understand that the S.R.C. and the M.R.C. have now set up co-ordinating bodies of their own to assist in providing computers for work financed by them; we welcome this move, and suggest that the other councils should do likewise.
238. We recommend that the Research Council laboratories should continue to have access to nearby University machines. This has often been the practice in the past, especially on the part of the Medical Research Council, but the arrangements have been informal ones which are likely to be jeopardised by internal University pressures. There is no doubt that several Research Council laboratories possess the experience to run their own machines, and growing usage will generate a further need. We consider, therefore, that as with the Universities each laboratory should eventually aim at possessing a computer of its own capable of meeting most of its normal demands, but that for problems of greater size or complexity it should have access as of right to the regional hierarchy based on the Universities. Although we make no specific proposal, we consider that a case might arise for a regional machine to be operated by a Research Council rather than a University; indeed, the Atlas Computing Laboratory of the S.R.C. at present operates as such on a national scale and is making a most important contribution to the computing strength of the Universities as a whole.
239. In addition to possessing their own laboratories, however, the Research Councils are empowered to make grants for special researches to be prosecuted in Universities but which for reasons of scale, urgency, or novelty are beyond the immediate resources of the U.G.C. Many of these grants require the use of computing at some stage, and it is desirable that this be done within the normal university framework wherever possible in order to encourage the interdisciplinary development of the art.
240. Nevertheless, the Science Research Council especially are receiving an increasing number of applications for computers to be operated on-line in very close electronic association with other equipment and in such a way that they would be very largely unavailable for any other purpose. They have also received applications for relatively small computers required for special research projects which would again make them unavailable for general purpose computing, and sometimes unsuitable for this latter purpose by virtue of their characteristics. Conversely, a central computer, although perfectly adequate for the general computing load of the university, may not be able to meet the special needs of a particular research project. In any of these cases we consider that the Research Council should itself meet the cost of the special computing facilities required by the research out of its normal budget. It follows that in deciding the grant to be awarded careful attention should be paid to whether special computing facilities are required, or whether the need can be met from existing installations. As far as fully on-line operation is concerned we have been able to make no distinction between the computer and any other piece of electronic equipment associated with the work; indeed, they may sometimes be interchangeable. For this reason it is not considered appropriate that on-line, including control computers costing less than about £50,000 should be subjected to review outside that applied to the grant for the research as a whole, although it would be helpful for administrative reasons if the Department of Education and Science were informed of such expenditure on a routine basis from time to time.
241. The characteristics of on-line and special purpose computers are dictated by the nature of the application and for this reason, as well as the ones already given, we do not attempt to evaluate the corresponding needs. Due allowance has already been made in the forecasts of the Research Council most concerned at the present time, namely the S.R.C., and as far as we can see the proposed scale of expenditure and the purposes to which the computers are to be put seem very reasonable and desirable. The typical cost of each such equipment is about £25-50,000, and according to the S.R.C. about £200,000 per annum may be involved over the next few years. Certain equipment required for film analysis in nuclear physics is substantially more expensive. It is important to bear in mind, however, that on-line equipment is usually required to feed its output into a larger computer for further analysis, and in specifying the equipment it is therefore essential to bear in mind compatibility with the larger machines to be used in this way.
242. However, we note that most of the equipment already agreed to or proposed is of United States manufacture, and we believe the attention of industry should be drawn to this, for we believe that the proper prosecution of research will throw up growing needs for on-line data processing not only in the physical and engineering sciences, but in many others, such as medical research, as well.
243. Before turning to the needs of the separate Research Councils we wish to point out that experience in the more technologically developed physical sciences has suggested that about 5 per cent of the annual budget of a research laboratory should be spent on the average on computing equipment of one sort and another. Of course, not all sciences require this level of expenditure, but it is reasonable to suppose that any research laboratory trying to make use of modern physical techniques for whatever ultimate purpose might find it profitable to set aside for computing equipment, depending on the nature of the work, between I and 5 per cent of its total annual budget. This range allows for the fact that the ratio of capital to total expenditure amongst the Research Councils varies from about one-twelfth to one-quarter.
244. The comments in Chapter II relating to peripheral equipment, software and staffing are mostly applicable to computing services operated by the Research Councils. Readers are again reminded that where specific computers are named, this is by way of indicating the general kind of computer recommended, and is, for example, without prejudice to proper tendering action.
Medical Research Council
245. The use of computers in medical and related biological research is still at an early stage of development. There is, however, clear evidence that scientists are rapidly beginning to realise their potentialities not only for increasing the speed and efficiency with which existing and well-defined manoeuvres can be undertaken, but also for making possible fresh approaches to problems which were hitherto insuperable and for opening up entirely new fields.
246. The Medical Research Council some three years ago examined the current use of computers for medical research and the consequent need to provide access to such apparatus within their own organisation. The outcome was a decision that certain relatively small computers were needed immediately for the sole use of Council staff working in the fields of biophysics and molecular biology, and two Elliott 803 machines and one Ferranti Argus were acquired for this purpose. These machines are being put to full-time use and have made possible considerable advances by the workers concerned. The conclusion was also reached that other staff requiring access to computers for a variety of purposes, but not their on-line use, could at that time have their needs satisfied by obtaining access to university machines. It was realised, however, that they needed assistance both in programming and in data preparation, and the Council therefore decided to set up a small Computer Services Centre equipped with programming staff and the necessary ancillary apparatus; this centre is in process of becoming fully operational. At this present time some use is being made of digital computers in about 75 per cent of the Council's establishments.
247. Provided that adequate facilities are made available in the Universities, it seems likely that in the immediate future the needs of Medical Research Council workers in centres outside the London area can be most efficiently met by their having access to university machines. This statement will, however, be true only if an adequate and efficient service is available in the university and if there is an obligation as of right to serve Research Council workers so that they may obtain a reasonably fast turn-round of their jobs. If this requirement cannot be met it would be necessary for the Council to consider the more expensive possibility of acquiring for their own purposes in a number of centres a part share of a large computer with a university, a regional hospital board or a commercial organisation. It might be mentioned in this context that a number of medical schools are already arranging for their own computing facilities apart from university provision, generally in conjunction with Boards of Governors of teaching hospitals.
248. In the London area some Council workers are at present enjoying extremely favourable treatment in the use of the University Atlas. It seems likely, however, from current estimates that in less than two years' time it will be impossible for M.R.C. staff to obtain adequate computing time or speed of service from the Atlas or from the recently installed IBM 7090 at Imperial College. Furthermore, in considering computer requirements for the London area, the growing needs of the National Institute for Medical Research and the eventual needs of the projected Clinical Research Centre have to be taken into account. Even if additional computer facilities are provided for the University of London, the number and size of Council establishments in this area make it desirable that the Council should acquire a computer of their own.
249. It is estimated that, in addition to the full-time use of computer facilities at present owned by the Council, M.R.C. staff will need the equivalent of some 6,000 hours computing time per annum on a KDF 9 over the next 2-3 years. During this time the Council wish to acquire a general purpose computer of roughly the size of a KDF 9 for the use of their staff in the London area, and to purchase a number of relatively small digital machines for use in laboratories studying real-time applications or handling large amounts of data. Additional provision of ancillary equipment and of staff trained in mathematics and statistics is also anticipated. While it is impossible to specify precise needs beyond the next 2-3 years, a considerable acceleration in demand after that time is to be expected.
250. We recommend that the M.R.C. be invited to consider acquiring computer equipment at the rate of about £0'2m per annum, and obtaining a KDF 9 for the London area. This machine, however, should be of the Culham standard, able to take programmes in advanced languages; we hope that it could be installed in 1966-67. We recommend that the M.R.C. Computer Services Centre should make their machine available to other users in or near London, i.e. that it should operate within the proposed London hierarchy.
Agricultural Research Council
251. The A.R.C. possesses an Orion I computer at Rothamsted Experimental Station, which is now inadequate. There are six research institutes in the Edinburgh area which also make use of it and which are now requesting facilities of their own.
252. The institutes in question, which have a total complement of about 200 scientific and experimental officers, are the Animal Breeding Research Organisation, the Poultry Research Centre, the Scottish Plant Breeding Station, the Scottish Station of the National Institute of Agricultural Engineering, the Moredun Institute (Animal Diseases Research Association) and the Hill Farming Research Organisation.
253. It has been proposed that the Orion should be upgraded and that the Edinburgh institutes should share in the use of a large machine at Edinburgh University: the latter has been proposed by a firm of consultants who have made a study on site.
254. We fully accept the proposals for Edinburgh. As discussed in Chapter V, we recommend a regional conversational arrangement serving the University, the Royal Observatory, and the A.R.C. Institutes; we have found no evidence that British manufacturers could undertake a project of this complexity to the required time scale.
255. It is clear that Rothamsted needs a larger and more modern machine allowing more sophisticated programming than is possible with the present machine. We consider that an Elliott 4130 should be installed in 1966-67. In our view extension of the existing machine would be at best an expensive short-term solution.
256. It is understood that the A.R.C. will shortly be setting up a Structural Chemistry Unit in London. Until the work of the unit has been clarified it is not clear whether or at what stage a computer should be provided; in the meantime we assume that essential needs can be met contractually with the University of London or one of its constituent colleges.
257. To cover the less well-defined and long term needs of the A.R.C. we have included a notional sum of £0.1m for the year 1967-68 and £0.2m thereafter, in the expectation that experience at Edinburgh, Rothamsted and London will generate a lively interest in computer applications. We fully expect that some of the Edinburgh institutes will find the need for some equipment of their own in addition to access to the regional machine.
Natural Environment Research Council
258. Owing to the fact that N.E.R.C. has only recently come into existence it has not yet been possible to make any realistic estimates of their likely requirements. The constituent parts of the Council make use of a great variety of university and other computers, but possess nothing of their own. They have use of the Liverpool, Glasgow, the National Physical Laboratory, and the Meteorological Office KDF 9 machines, and if all these machines are upgraded to Culham standard the facilities available to them will be greatly improved in quality as well as quantity.
259. While for some time to come the Council will presumably continue to obtain time elsewhere, we strongly recommend that they encourage applications for small computers from one or two of their establishments, such as the National Institute of Oceanography, and we have included a sum of £0.3m in the year 1967-68 with this in mind.
Science Research Council
260. The S.R.C. is unique in possessing two very large nuclear physics laboratories, created under the National Institute for Research in Nuclear Science, which are operated for and largely by the Universities as central facilities. All of the work of these two laboratories is concerned with large accelerators which cannot be used effectively except in close association with correspondingly large computers operating, in part, on-line to the accelerators, either directly or through small control computers. The computing requirements of nuclear physics were recently examined by a joint working party of the former D.S.I.R. and N.I.R.N.S., and it was recommended that both laboratories should acquire large central computers of their own.
261. In the case of Daresbury the issue is clear because there is no other computer available to them for close, constant and reliable operation with the accelerator. A wealth of experience exists to show that in work of this kind computer demand doubles each year and a large expandable machine is therefore essential. There is no British machine of the required characteristics available at the right time, and the application is for an IBM 360/50 together with on-line control computers. We support the application and recommend that the order be placed at once, but envisage that this machine may require upgrading to 360/65 standard within two or three years.
262. In the case of Rutherford High Energy Laboratory the issue might appear not quite as clear because of the presence of the Chilton Atlas nearby of which the Rutherford Laboratory already has one-quarter of the use in addition to its own much smaller Orion. The joint working party examined the technical possibility of using Atlas rather than purchasing a further machine. However, the estimated demand exceeds the total capacity of Atlas especially when allowance is made for the fact that in the years 1966 and 1967 there will be a most serious shortage of computer time for the university film analysis groups most of which are involved in the international programme of C.E.R.N. This shortage alone amounts to at least one-third of Atlas and the Rutherford Laboratory have taken the responsibility for meeting this demand. Further the Chilton Atlas is making a vital contribution to general university computing, and that of certain Government establishments, the need for which cannot possibly decline before the programme of new provision recommended in this Report is at least in its third year. The Rutherford Laboratory must therefore have a machine of its own.
263. No British machine is available to meet the specifications, especially of size. The choice lay between the CDC 6600 and an IBM 360 machine. The necessity for on-line use and the desirable feature of full-time sharing led them to the IBM 360/75. We support the application, but regard the final choice between these machines as a matter within the discretion of the S.R.C.
264. The S.R.C. wishes to hire both machines, for reasons of flexibility. We welcome this because in addition it would make possible their replacement by British machines if and when suitable ones appear on the market.
265. The S.R.C. is also unique in having inherited the largest single computing installation in the country. The Chilton Atlas, costing about £3.5m, was obtained to serve the needs of Harwell, the Rutherford Laboratory, the Universities and Government establishments. Of these, the first two bodies each have one-quarter use at the present time. We understand that Harwell is to obtain its own machine and we have recommended that the Rutherford Laboratory should do likewise. Within 1 to 2 years substantially more time will therefore be available to the remaining users, of which the greatest share is expected to go to meet University demand.
266. The Chilton installation is an extremely expensive one. It is performing reasonably well, and the full value of the investment should be obtained. Its useful life will continue beyond the period of our Report. For a few years the national service which it provides will be essential. Thereafter it could become a regional machine serving the Universities of Oxford, Reading, and perhaps Bristol and Birmingham; or it could be replaced by a more powerful machine, presumably the most powerful available, and continue to give a national service over and above that provided by the regional hierarchy; technical developments alone make any decisions at this time difficult. The proper development of this laboratory, however, requires the continual purchase of substantial items of equipment at a rate of perhaps £0.2m per annum. We have included this figure; in addition we have included a purely notional sum of £2m in the last year in case the S.R.C. should find it necessary to replace Atlas with a more powerful machine. We also recommend that a time-limit should be set beyond which inefficient languages should not be accepted at Chilton.
267. Before leaving the Chilton Atlas we feel bound to comment on a matter we regard as very important. It is clear from our proposals for the two former N.I.R.N.S. laboratories that nuclear physics has been allocated a very large share of the country's computing power. We do not consider that the nuclear physicists have exaggerated their requirements; on the contrary, they were amongst the first, owing to the nature of their work, to appreciate the large growth rate of modern computing, and they have done no more than to make proper allowance for it. It follows, however, that their needs have inevitably been over-provided for in the first year or so. That being so, we recommend that the S.R.C. should consider it to be entirely within their discretion to allocate a substantial fraction of the time of either or both of these powerful machines to other purposes during the next two years, the period during which excess University pressures are likely to be greatest. If they were so to decide, machine time could probably be allocated most satisfactorily through the auspices of their Atlas Computer Laboratory at Chilton, although additional operating staff might then be required.
268. We now turn to establishments of the S.R.C. outside the former N.I.R.N.S. framework. The Royal Greenwich Observatory at Herstmonceux has recently ordered an ICT 1909 machine. Their statement that it will be adequate for five years may in the end prove to be optimistic. We have already noted that Herstmonceux is close to the University of Sussex who also wish to have an ICT 1900 series machine. There should be close collaboration between the two installations.
269. The Royal Observatory at Edinburgh has access to a number of distant computers including the Atlas machines at Manchester and Chilton. They require a small machine for on-line use, but our main recommendation is that for general purpose computing they should be included in the conversational project centred on Edinburgh University.
270. The Radio Research Station at Slough operate a very small Pegasus belonging to the Road Research Laboratory and has applied for an ICT 1909. There is a genuine need for a small computer well supplied with peripheral equipment, capable of expansion and compatible with larger nearby machines. The small computer on the premises is required for most of those individual theoretical problems which require the use of a computer, the analysis of comparatively small amounts of data from several different experiments, and the processing of satellite positional information received in the morning in time for distribution by telephone to workers who will use it in their evening observations. Computations of these kinds require the close attention of those responsible for them and can be done efficiently only with a computer on the site. We recommend that their application be approved.
271. The S.R.C. is also responsible for the Space Research Management Unit which provides technical management for the U.K. civil space research programme. Where computing is concerned its responsibilities are to ensure that adequate time is available on university computers, and others where necessary. The proposals for regional organisation of university machines should make its work easier in this respect, especially in the London area, as should its recent acquisition of the former N.I.R.N.S. laboratories.
272. There are finally two S.R.C. activities of a less well defined nature at the present time. It will be responsible for the execution of the Fleck Report on Radioastronomy if that Report is accepted. We merely wish to point out that a £4m programme of radioastronomy will imply a correspondingly large computing load the extent of which was not estimated in the Report. We hope that this will be rectified before the Report is considered in detail, for it would not surprise us if the programme were to imply computing equipment costing around £0'4m in addition to the usual on-line control computers. The S.R.C. is also dealing with an application for a grant for an Institute of Theoretical Astronomy which would be likely to ask for its own computer.
Summary of Recommended Research Council Expenditure
273. The cost of the recommendations of the Working Group for Research Council's hardware are tabulated below:
274. As in Chapter V, 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.
275. Of the £8.45m some £4.6 m represents American equipment, on the basis of outright purchase.
276. Most of our recommendations for Research Councils have high priority within the programmes of the Councils.