Flowers Report

Jump To Main Content

Jump Over Banner

Home

ACLLiteratureManualsAtlas Autocode

Jump Over Left Menu

Chapter III Comments on Applications Received by the University Grants Committee

87. In this chapter we make comments on the needs of some of the Universities where several problems of general interest have arisen. Our information comes partly from the visits we have made, partly from the detailed correspondence held by the U.G.C. and partly from our own knowledge. We describe the situation in London at some length: this University consists of a number of semi-autonomous colleges of varying size, so that much may be learnt from them of more general application.

London

88. The demand for computer time in London has increased tremendously during the past few years. When their Atlas was ordered as a replacement for the earlier Mercury, the University was to have a quarter share of machine time and a commercial company a further quarter; the remaining half was to be sold commercially in order to pay the remaining costs of the machine. In fact Atlas has turned out to be only about half as efficient as expected. Even had this not been the case, however, the University states that it seriously underestimated the growth in demand. Consequently there has been an acute shortage of machine time throughout the University and many research programmes have been seriously delayed. There is now a clear case for a much larger central facility the purpose of which is to provide a regular and efficient general service.

89. The need for multiple-access facilities was mentioned especially in the field of engineering design and we were advised that this mode of operation might become widespread in the future. The University also stressed the need for trained computer management personnel and programmers and expressed an interest in helping to meet the national shortage.

90. The difficulties experiences by the major London colleges in getting access to the Institute Atlas has highlighted a demand for local machines such as the IBM 7090 recently installed at Imperial College and already fully loaded. Some idea of the need at University College, for example, is clear from their statement that they need 90 hours per week 7090 time (the best part of the whole machine) by 1966. The heavy emphasis placed there upon postgraduate training and research, especially in nuclear physics, space physics, and chemistry, will continue to generate demands for large scale computing.

91. Similarly, at Queen Mary College the Physics Department's experimental activities are resulting in the need for large scale analysis of data requiring readily available machine time during the whole of the working week. Frequently the need for on-line working has cropped up, for example in crystallography at Birkbeck College. Other colleges like Bedford, Westfield and Queen Elizabeth have computer work which they would like to do were facilities available. At one of these colleges, some idea of the lack of suitable facilities is indicated by the action of one man who found it quicker to do his computing work in Canada, flying there and back at frequent intervals.

92. There are important developments in the London University Medical Schools. Statistical analysis of patients' records and diagnosis cannot really get going until proper facilities are made available. In this field there is an urgent need for compatibility decisions to be taken in conjunction with any central computer facility. It is also expected that the various departments of economics, especially at University College and the London School of Economics, will have steeply increasing requirements.

93. All the colleges, even the smaller ones, wish to teach computing methods to undergraduates and to have quick turn round for program development. It is difficult to see how this could be met unless each college possessed at least a small machine of its own.

94. In addition to the constituent parts of the University it is important to bear in mind other research laboratories in London, such as those of the Medical Research Council, who currently make use of University facilities on a somewhat informal basis.

95. A careful estimate of the needs in the University of London up to the year 1970 has been prepared for us by the Committee of Management of the Institute of Computer Science. They based their findings on:

  1. A knowledge of the present distribution of computing load on Atlas, together with estimates of computing done on other machines.
  2. Individual estimates by colleges and schools of the University of their requirements now and in future years.
  3. Overall estimates of the growth rate to be expected.

Present Distribution

96. From statistics compiled during the period June 1964 to March 1965, the distribution of the total computing load of the University at present is approximately as follows:

%
Imperial College 32.0
University College 26.0
King's College, Q.M.C., L.S.E., Birkbeck and Royal Holloway College 27.5
Institute of Computer Science 7.5
Other colleges, institutes, etc. 7.0

Individual Estimates

97. (a) The largest users (Imperial College and University College). It is proposed to use an Atlas unit as a measure of load, this representing 120 hours per week of useful computing, with technical and operating efficiency of Atlas about 25 per cent better than now. On this basis estimates for the two major users are as follows:

1964-65 1965-66
Atlas Units
Imperial College 0.16 0.48
University College 0.13 (0.25)
0.29 0.73

Bubble chamber analysis is excluded from the Imperial College estimates, which have been scaled down by a factor of 0.8 (to allow for expected increased Atlas efficiency). University College estimates a load expressed as about 90 hours of 7090 time per week by the end of 1966; this is roughly 0.24 A. units, and seems low in relation to the I.C. estimate if the current ratio of 4:5 were maintained. However, despite all the uncertainties involved, a total for the two colleges of 0.75 A. units in 1965-66 seems unlikely to be an underestimate though a range of 0.6-0,9 units is possible.

(b) Other major users. Some colleges have estimated their needs for 1965-66 and 1969-70 in terms of Atlas hours per year:

1964-65 1965-66
Atlas hours
Queen Mary College 400 3,000
Royal Holloway College 400 1,200
Birkbeck College 210 725

King's College have given no overall estimate, but on the basis of present use 400 hours in 1965-66 might also be expected. L.S.E.'s needs are also not closely defined, but an estimate of 700-1,000 hours in 1969-70 seems likely. On the basis of these estimates, these five colleges might require about 9,000 Atlas hours in 1969-70, or 1.5 Atlas units.

(c) Smaller Colleges and Schools. Several colleges and schools (such as Queen Elizabeth, Wye, Oriental and African Studies, Pharmacy) have estimated requirements varying from 35 to 55 hours in 1969-70. Similar needs, or perhaps rather greater, may be expected from Bedford and Westfield Colleges. A total of 500 hours (0,1 Atlas unit) in 1969-70 would probably cover institutions in this category.

(d) Medical Schools and Institutes. From estimates made by seven medical schools, a total of 140-210 Atlas hours in 1969-70 is obtained. Some Institutes also forecast small requirements of about 10 hours/year, but larger estimates are given by Institutes of Cancer Research, Psychiatry, Child Health and the Postgraduate Medical Schools which might add up to 1,000-1,500 hours in 1969-70. There are many medical requirements which would not easily be satisfied on Atlas at all.

A total of 1,500-2,000 hours (0.25-0.33 Atlas units) in this category seems reasonable.

Growth Rate

98. An annual growth factor of 2 is commonly applied. This has been exceeded in London in the past year, and Imperial College has assumed a factor 3 in going from 1964-65 to 1965-66. Thereafter smaller increases may be appropriate, and the following are assumed very tentatively (based on Imperial College documents):

1965-66 1966-67 1967-68 1969-69 1969-70
1.0 2.0 3.5 5.7 8.3

Present indications are that about 60 per cent of the total University computing load arises from Imperial College and University College, and that in 1965-66 this will amount to about 0.75 Atlas units.

Combining this evidence with the growth figures above we obtain the following estimates for the years:

(Atlas Units) 1965-66 1966-67 1967-68 1969-69 1969-70
I.C. and U.C. 0.75 1.5 2.6 4.3 6.2
University 1.25 2.5 4.4 7.1 10.4

A partial check on the extrapolation may be obtained from the estimate given by other parts of the University for 1969-70. These are as follows, on Atlas units:

Five major colleges 1.5
Smaller colleges and schools 0.1
Medical schools and institutes 0.3
Institute of Computer Science 0.8
Total 2.7

99. The estimate of 2.7 should be compared with 10.4-6.2=4.2 in the previous table. There is a discrepancy which may be due to some omissions (e.g. work done for Medical Research Council and other institutions in the London area), to underestimates by some colleges, or to too steep a rise being assumed for I.C. and U.C.

100. If it is accepted that the basic figure for I.C. and U.C. in 1965-66 will lie between 0.6 and 0.9 then it seems safe to conclude that the total university load for 1969-70 will, if facilities are not unduly restricted in the meantime, lie between 8.3 and 12.5 Atlas units.

Cambridge

101. This University has been one of the pioneers in the development of large fast computers. The present Atlas-Titan installation, in part designed locally, is providing a better service than most other Universities at present enjoy, although criticism from users is not entirely absent. The University computer committee states that it wishes to install a new large machine in two or three years time although there has been no formal application to the U.G.C. They have recently acquired in exchange for machine time a £75,000 disc file from I.C.T. They wish to add a further 32 K of core storage and a number of additional peripherals at an estimated cost of £175,000, in order to improve their general service. They have informally expressed interest in experimenting with a multiple-access system.

Manchester

102. This University, through the pioneering efforts of its Department of Electrical Engineering, has been able to provide its own central computer since 1951 without appreciable U.G.C. assistance. The present Atlas is already fully loaded and accommodation for it is totally inadequate. There are several major users within the University, the Physics Department for example, and a service is also provided for the Manchester College of Science and Technology (the Faculty of Technology of the University), who possess no machine of their own, in addition to many other universities amongst which Edinburgh and Nottingham are connected by data link. The present machine is owned by I.C.T. with whom its use is shared and the contract runs out in 1970. The University has little chance of installing a new machine before 1969, by which time it will be grossly overloaded, but it must do so before 1970 and the building programme has been modified accordingly. At that time they wish to have a substantially larger machine, and in the meantime because of their small core store they have asked for a disc store to increase the efficiency and flexibility of their present installation. Space allows of no other immediate improvements in hardware. It is intended to separate computer service from the research activities of the Department of Computer Science when the new building is available.

Edinburgh

103. The University claim in a document dated June 1964 that their share of the Manchester Atlas, amounting to 15 minutes per day, has never been obtained due, so it is said, to the malfunctioning of the computer and its software. The situation has improved somewhat since then but the feeling is that the link is going nowhere to meet the University need. According to Manchester, however, few attempts to use the link have been made in the last 12 months. Whatever is the case, it must be said that Nottingham, who have had the same sort of link, managed quite well. The University of Edinburgh clearly contains a body of active computer users; one example is the group associated with the English Language Research Unit. Two of these have commendably put aside their work for six months to join the computer people at Edinburgh in writing an Atlas Autocode compiler for use on the Glasgow KDF 9. Other examples of major cases of need in the University are in the Department of Astronomy for telescopic control, and analysis of data from automatic measuring devices and stellar spectra. In the Royal Infirmary the cardio-respiratory unit is arranging to make records of measurements on patients' blood pressure, pulse rate, respiratory rate and so on. One marked need for the computer analysis of data is at the International Seismological Centre which is to act as a centre for the distribution and analysis of data from seismological stations all over the world. The need in Edinburgh University is supported by the activities of several Research Council laboratories nearby, including the Royal Observatory.

The KDF 9 Universities

104. During the past year a substantial part of the computing at British Universities has been borne by the first round KDF 9 installations at Oxford, Leeds, Glasgow, Birmingham, Liverpool, Newcastle and Salford. A further KDF 9 is to be installed at Nottingham in December, 1965. The first round grants were not uniform but varied between £0.150m at Oxford and £0.225m at Glasgow. The grants were all made during 1961 but deliveries did not start until late in 1963 and have only just been completed. This delay in delivery has contributed in good measure to the present weakness in the computer state of Universities generally. One major requirement of the KDF 9 system was to take up the load on the Mercury computers which had been the backbone of University computing before 1963. Many programs were written in Mercury Autocode and no compiler was made available by the manufacturers. It has since been produced by I.C.T.

105. The Algol compilers produced by the manufacturers were written for the smaller core stores (8K and 16K) recommended in the first round: no real improvement in Algol can be achieved by adding core store unless a new compiler is forthcoming. All the KDF 9 installations run one shift and some two shifts; but here again difficulties are met with. The financial position at some Universities is such that more computer time cannot be provided because extra maintenance is not possible, and, for example, Oxford run two shifts but maintain the computer for one only. This sort of working is feasible on very reliable machines only.

106. Since the first round grants some additions have been recommended by the U.G.C. The nature of the additions give a clear indication of the difficult conditions which have been met at the KDF 9 installations. For example, at Liverpool £90,000 has been made available to augment a 3 module core store (12 K), 3 magnetic tape decks and a punched tape orientated, line-printer machine for the following very good reasons:-

  1. One tape deck is used for output, and one for control and library programs. This leaves one tape deck for the programmers. The break down of one tape deck (not an infrequent occurrence) effectively immobilises the whole machine.
  2. The Mercury Autocode compiler which has become available needs at least 16 K of core store and is therefore not usable.
  3. The Tidal Institute cannot use the punched paper tape input/output system because their data is essentially punched card orientated.

107. These examples at one University highlight the general need for more magnetic tape decks, more core storage, and card input/output.

108. The situation at Liverpool is typical of all the present KDF 9 installations. There still, however, remains the software problem which cannot properly be resolved until enough core storage is available at all the installations to provide properly efficient compilers for the major languages, Algol and Fortran.

109. The Working Group estimates that because the initial configurations were too small and because of subsequent hardware and software difficulties, the KDF 9 installations have not markedly improved the overall state of these Universities since the Mercury days; this is to say that since 1961 there has been a period of painful development which compares very unfavourably with developments in Europe and the U.S.A. where average figures of annual expansions are between 1.5 and 2'0 in computer power. The population of the KDF 9 Universities borders on 50,000 and they should have been a rich source of computer-trained academics who are now in such short supply.

110. In considering the work which has been held up because of the lack of proper facilities at the KDF 9 Universities, the Working Group finds that in the fields of economics, psychology, linguistic analysis and in some branches of theoretical physics where punched card equipment is vital, there has been almost complete blockage. In the fields of crystallography, education, and the technological applications of computer techniques, there have been serious delays.

111. The needs for the typical KDF 9 installations have been clearly expressed by Birmingham. In hardware, they now require readers and punches for both cards and tape, line printer, 32K store, graph plotter, more tape units; and in software a proper monitor system, compilers for Fortran, Algol, Mercury Autocode and LISP type languages.

112. There are two KDF 9's in the Atomic Energy Authority which are provided with all these facilities except for the LISP compiler: one at Winfrith, the other at the Culham Laboratory.

The University of Wales

113. For nearly all purposes relevant to this enquiry the four Colleges of the University are effectively separate institutions. The present equipment is as follows :-

Cardiff Stantec Zebra
Swansea IBM 1620
Aberystwyth IBM 1620
Bangor Elliott 803

114. Some very minor additions have been made to the Bangor machine only in the past year. All four colleges have recently expressed their needs for new computer facilities and one college, Aberystwyth, has had a U.G.C. application for a major American installation deferred. In this connection it is sometimes suggested that it may be economically preferable to install one large machine for the benefit of all the colleges. We do not believe this is a practicable solution.

115. All the present machines are fully loaded; the Aberystwyth computer, for example, runs a 24 hour day, the majority of the machine time being used for research in mathematics and the physical sciences. Some, but by no means enough, undergraduate instruction is undertaken on the IBM 1620, and the College is anxious to run a postgraduate course in data processing and systems analysis in 1966. This will not really be possible without a bigger machine. The College's work on the solution of partial differential equations encountered in the study of the Earth's atmosphere is much too extensive to be handled by their present system. Once more the need for card input and output is highlighted at Aberystwyth by the use which the Plant Breeding Station is trying to make of the 1620; the introduction of a Faculty of Social Studies and the expansion of the Economics Department will demand computer time on a punched card-magnetic tape computer.

116. Similarly, the growth of the University College of Swansea is such that numbers are expected to reach a total of 3,000 students in 1967-1968 and 4,000 in 1972-1973. The greater part of their computing already has to be done elsewhere at a variety of machines. However, a considerable reorganisation and expansion of staff would be necessary before they could operate an efficient service. Much development of research is already in progress, however, and the Departments of Pure and Applied Mathematics, Physics, Engineering, Chemical Engineering and Metallurgy are all expecting to move from temporary quarters into new permanent buildings with greatly improved facilities during 1965-66. The College expect a considerable expansion of research activity in this group of computer conscious departments.

117. The use of data links to relieve some of the current pressure for computer time has been examined in detail. It is estimated that with the best available transmission system it would take approximately ten hours daily to transmit the college's total requirements to a remote computer.

118. The Bangor machine which was installed in 1964 is not being used to its full capacity because of a shortage of ancillary equipment, but this has now been rectified by the U.G.C. and the College is to report on the present installation and their future requirements shortly. Bangor is one college which is not suffering acutely from shortage of computer time at this point in time, although the present installation will probably soon become inadequate for many research projects in the expanding science departments.

119. The least adequate computer in the University is the Stantec Zebra at Cardiff, the biggest college. The computing needs typified by the large crystallographic group in the Physics Department of the College are for a 16K, 5 magnetic tape, paper tape and card input/output machine. A line printer and graph plotter are considered essential peripherals.

120. The general position in the University is that three of the college computers are inadequate for present needs and one soon will be. The difficulty of using a large central installation is pointed out by one college and another makes the important point that no machinery exists whereby an application from the University as a whole could easily be made. Further difficulties are that Swansea will shortly be accessible from Aberystwyth only by private car, and that requires 2i hours driving over mountain roads. However, there appears to be a need for Aberystwyth and Bangor to cooperate in the computer field and for a similar join-up between Cardiff and Swansea; in both cases the computer interests are similar.

The New Universities

121. The problem in the computer field facing the new Universities is that it is very difficult for them to estimate their requirements for the five year period with which the present survey is particularly concerned. The new Universities are very different in character; some are clearly scientifically and industrially orientated and computer conscious, while others are much less so. Moreover, their requirements can be expected to depend rather sensitively upon the needs of relatively few people for some time to come.

122. Probably the less computer conscious of the new Universities have requirements which lie hidden at this point in time but which may develop very rapidly indeed in the non-physical sciences. It has been the experience of the Working Group, (and this is of particular interest to the new Universities), that while the application of computers to scientific problems has been a spur to the progress of research into physical science generally, there is a vast field of progress in medicine, law, economics, business administration, banking and commerce which is ripe for computer development. This field has not yet properly been exploited even in the U.S.A. where computers are accepted as essential equipment, the reason being that a good deal of intelligent application of computer techniques, coupled with an application of something better than the obvious, is required in these areas.

123. In many ways the new British Universities are in a very good position to accept the challenges of modern computer developments. They are well placed to decide on the fields of new computer application for which they have special qualifications, and are free to plan the sort of computer facility, properly housed with an eye to future expansion, to take advantage of the undoubted knack British academics have for the innovation of advanced techniques. This is not to say, however, that all the new Universities should plan very large computer centres, rather that they should plan computer development to meet the needs of the academic and research staff who could make most rapid strides in the field of computer application.

Belfast

124. The University's present local computer is a Deuce loaned to them by the U.S. Navy. The University teaching programme in numerical analysis, statistics and automatic computation has been expanding rapidly and is operating at a high level of activity. There are at present some seventy students in attendance at one or more of the courses. A substantial part of one of the undergraduate courses in Mathematics provides for numerical analysis and automatic computation.

125. It is estimated that the research needs of the University expressed in terms of the ICT 1905 computer ordered by the Northern Ireland Government amount to 80 hours per week by the end of 1965, plus 300 hours per year on the Chilton Atlas. They are already one of the largest users of Chilton and the University also used 600 hours IBM 7090 time in the U.S.A. during the six months preceding our visit. The Working Group found a number of very well informed users in the University having between them experience of a wide range of computing machines and a large number of computing and data processing organisations, so that it can be said that the University is already familiar with the organisation and use of a modern computer.

126. The Deuce runs a 24 hour, 7 day week and there is an increasing backlog of material awaiting processing. This has led some people to seek time elsewhere and others to cancel their projected researches. The outstanding work load comes from the Department of Applied Mathematics which has a large and active graduate school whose interests lie in the fields of atomic and molecular physics, the physics of planetary atmospheres, theoretical chemistry and in associated numerical problems. The Diploma Course in numerical analysis which makes demands on computer time was taken by 10 students in 1963-64 and the numbers are increasing. The University feel a great need for a proper local machine facility in connection with this work.

127. The unique position of Northern Ireland for certain kinds of economic studies is important to the computer needs at Belfast. The population is of a convenient size, relatively stable with a natural boundary, and there is a small number of relevant authorities with whom communication is easy. The Departments of Medical Statistics and of Medicine, the Department of Education and Public and Comparative Law are all eager to exploit this situation but access to a large store computer is essential.

Aberdeen

128. This University has a fully loaded Elliott 803 machine, and is chosen for discussion because their application for equipment has been deferred by the U.G.C. Panel since November 1964: their case is urgent.

129. An essential feature of the University's proposals is that the system should meet the needs of the Faculty of Medicine. These involve fairly early provision of time-shared on-line consoles located within the hospital complex. Plans for work in the Department of Engineering have also progressed to the point at which the desirability of operation from separate consoles is clear. The University makes the point, that such facilities are not available on any existing British machine in the price range which the University has been examining.

130. The multi-console facility is the primary reason for their choice, but there are others which the University considers are important in reaching a decision. Briefly these are:-

131. There are four Government supported research establishments in or near Aberdeen; these are the Macaulay Institute for Soil Research, the Marine Laboratory, the Rowett Research Institute and the Torry Research Station. There are M.R.C. and A.R.C. Research Units within the University. The Elliott 803 machine of which all have made some use cannot help further and all are increasingly conscious of how their work can be aided by a good local computer.

Lougbborough College of Technology

132. The IBM 1620 which since 1961 has provided the computing power, has been used to acquaint undergraduates with the use of computer techniques with good effect. The college has a good body of staff skilled as programmers. The computer time is divided between teaching staff and research staff and has been running at well over 120 hours per week since October 1963. Computer usage doubled in the first nine months of 1963 and it must be accepted that since October 1963 no further computer time has been available. A recent forecast of future requirements in the College shows that between 1965 and 1970 their usage will roughly double each year.

133. The stated needs of the College for a further computer are of general interest. In summary:

Hardware

  1. Increased speed to handle forecast workload of 100 IBM 1620's.
  2. Increased throughout by use of multiprogramming and autonomous peripherals.
  3. Sufficient core storage to assure proper use of a monitor system.
  4. Sufficient backing store.
    1. Random access to permit processing of non-sequential data.
    2. (
    3. Magnetic tape to facilitate rapid storage of large blocks of data.
  5. Peripherals, particularly a fast line-printer, capable of handling the amount of input-output resulting from use for mainly scientific purposes.
  6. Additional expansibility by possible interchange of central processors, implying availability of larger compatible machines.
  7. Provision of multiplexor channel with the possibility of outstations.

Software

  1. Availability of a supervisor program providing some multiprogramming facilities.
  2. Availability of an assembler and a compiler.
  3. Development of both Algol and Fortran compilers.
  4. Development of a full 'supervisor' system with clock-controlled interrupts, dynamic allocation of storage, program error correction and updating, batch processing from magnetic tape, etc.
  5. Provision of further software, packages, including matrix operation, critical path, linear programming and Cobol.

134. The work for which the new computer is required is typified by the future research work of the Department of Civil Engineering concerning the interactive solution of matrix problems in engineering structures, critical path network analysis related to cost analysis and budgeting. The electrochemical research work in the Departments of Applied Chemistry will demand much computer time. The Computer Centre planned on closed shop lines has need for time for its own research on compiler-compiler programs, timetable organisation, practical teaching, finance, student records, and library data.

The Teaching Hospitals

135. The Working Group took especial pains to acquaint themselves with the needs of the teaching hospitals and with medical applications of computers generally. We took evidence from medical representatives at Belfast and Manchester, and also met representatives of about a dozen medical schools in London. We found almost complete unanimity of view. In a later Chapter report on our evidence from the Medical Research Council.

136. The work of the teaching hospitals which requires the support of computer techniques varies from quite elementary data listing and sorting of a kind generally associated with commercial applications to the quite sophisticated numerical calculations of the scientific user. Some computer work is already undertaken in a number of teaching hospitals, usually the more scientific work because of the present difficulty of handling the very large amounts of data associated with routine medical work.

137. As expressed by an experimental neurologist: The difficulty is not so much the computing time we require but the problem of getting the very large amount of data which arises from biological experiments into a computer. When a sufficiently pointed question can be asked it is usually possible to carry out a rather massive process of data reduction locally, either at the time of the experiment or shortly after it. To do this one needs locally some equipment for data reduction.

138. The needs of a research psychologist are similarly expressed: What is needed is a data link to a time sharing computer, or extensive access to a medium scale machine of our own located on the premises. It is clear that a computer system using punched paper tape is quite unsuitable for the kind of work needing to be done by the teaching hospitals. For example, patient histories need to be updated regularly, at times daily or even hourly. The details of a treatment consisting of drug dosage at frequent intervals can and should be logged automatically in order to prevent possible error and in order to relate the effect of the treatment to the patient's past and future medical state. This sort of work must be done using punched cards and magnetic tape.

139. An Institute for Child Health for example has the following specific needs :-

  1. Statistical handling of large numbers of punched card data, recording a variety of measurements. Print outs are frequently required and statistical calculations often of a sophisticated nature such as canonical variates and multiple discriminant analyses.
  2. An on-line facility for linking with a flying spot shape sensor to be used for measuring changes in shape in bones and to count automatically changes in central nervous system elements during brain growth.
  3. An on-line facility for dealing with neurophysiological events in experiments on the growth and behaviour of young animals with indwelling electrodes.
  4. A random access facility for data storage and the retrieval of hospital records.

140. These requirements highlight the essentials of a particular teaching hospital computer facility which briefly seem to be:-

  1. Punched card, magnetic tape input/output.
  2. Access to a local data processing central unit.
  3. Access to a larger facility for large scale data processing.

141. Another teaching hospital is making an analysis of second generation obstetrical records which could result in valuable correlations between parent illness history and childbirth condition.

142. The work on encephalography is still in its infancy. At this time the encephalograph has a limited use because the data it produces cannot be assessed except by the eye of an experienced neurophysiologist. Yet the data must contain a lot of critical information about brain damage and the location of a brain disorder which could only be revealed by a statistical breakdown of the trace and an acute analysis of the information content of the waveform. Rapid strides in brain surgery, it is claimed, could be made by the application of computer techniques to this special field.

143. The enormous potential of computers in the medical and biological fields is only beginning to be appreciated. It may well be that in terms of volume of data processing this work may outstrip the physical sciences very rapidly in the coming decade.