An announcement is expected shortly from the Science Research Council that the Rutherford High Energy Laboratory, at Chilton, Berks, is to be equipped with one of the biggest data processing complexes now available - an IBM 360/75 worth at least £1M.
The Laboratory houses Britain's Nimrod atom smasher, third largest in the world, and the problems it is having to tackle in fundamental nuclear physics have made the provision of a very large system essential. This is expected to be available in about two years.
The Rutherford system will be the first of the very large next generation IBM machines in Britain. It is believed that negotiations are in progress for two more possible users being the Central Electricity Generating Board and Rolls-Royce.
Meanwhile there is a good deal of speculation in company circles on the type of computer which will be chosen for the third of the major university computing centres envisaged by the Flowers Committee.
The report, endorsed in its major points by the Government just before the recess, provided roughly for a doubling of the money to be spent on university computers over the next six years from the £10M originally earmarked.
The three centres are London, Manchester and Edinburgh.
The performance of the new £250 000 KDF computer is causing disappointment at Oxford and other universities where it has been installed. Why, and what is to be done about it is explained below by our Science Correspondent, DAVID PERMAN.
Hopes were running high among Oxford research workers last year, when the university sold its six-year-old Mercury computer and began operating the new £250,000 KDF 9.
But many of these research workers have been disappointed, and the reason for their disappointment goes much deeper than the teething troubles that most new machinery produces.
Similar disappointment has been felt at the other six universities which received the English - Electric - Leo KDF 9 computers.
What went wrong has now been noted by the Flowers Committee, whose report on computers for research was published last Friday. They recommend that all seven KDF 9s should be upgraded as a matter of extreme urgency. The upgrading of the Oxford computer, which is housed in the former Engineering Laboratory at the top of Parks Road will take place between July this year and July, 1967.
"What the Flowers Committee says about the KDF9 computers is absolutely true," says Prof. Leslie Fox, Director of the Oxford University Computing Laboratory. They may have exaggerated a little, but the new computer has proved only five or six times better than the Mercury, rather than the 30 times that we had expected.
There are two main ways in which the KDF9 installations have proved inadequate. They lack the peripheral equipment such as would make them able to deal with punched cards, that is vital for many fields of non-numerical research. And the KDF9s are not yet able to deal efficiently with the two main international languages in computing - Algol and Fortran.
The Flowers committee report says that there is almost complete blockage in the fields of research where punched card equipment is necessary - such as economics, psychology, linguistic analysis and some branches of theoretical physics. In crystallography, education and the technological applications of computer techniques, there have been serious delays, says the report.
Prof. Fox agrees that this is true of Oxford. We have had to suspend virtually all research on non-numerical problems, and we have not had time to allow undergraduates near the machine, he says.
The model for upgrading the KDF9 computers will be the KDF9 at the Atomic Energy Authority's Culham Laboratory. Since January last year this machine, which has all the required peripheral equipment and will probably get more has been working on a special programming system developed by the Laboratory and the manufacturers together.
But upgrading will not solve all of Oxford University's computing needs. More staff is needed, particularly for maintenance work. The Flowers Committee, in fact, singles out Oxford as a university where the computer is run on two shifts but maintained for only one.
The cost of upgrading will be met by earmarked grants from the Government. But for more staff, increased maintenance and, perhaps, new buildings for the computing Laboratory, the message contained in the Flowers Committee report will have to be recognised by the University Grants Committee and Oxford University.
The report further recommends that Oxford should be given priority for a replacement to the KDF9 in 1970.
This is gratifying, says Prof. Fox. It could mean one of two things - either that the committee recognises the vast amount of computing needed in Oxford or that we have made better use of the machine we have got. I should like to think both are true.
Oxford University is a large user of computing time. It uses as much time as any other university on the Atlas computer at Chilton, near Harwell, and the IBM machine at Imperial, College London. Oxford's nuclear physicists do some of their computing abroad at Geneva.
But Oxford entered the computing field later than many other universities. Before 1959, when the Mercury computer began working in a house in South Parks Road, there was practically no computing in the university. The Mercury did 36,500 hours of computing and cost £75.000. It was sold last year to a Birmingham firm for £450.
The longer-term needs of Oxford are recognised by the Flowers Committee in its study of the national need for very large machines. The main recommendations of the report. which the Government already has accepted, are for computing centres to be set up at Cambridge, Manchester and Edinburgh. It also recommends large machines, probably American ones on hire, for the Rutherford High Energy Laboratory at Chilton and for Harwell. All of this means that the Atlas Computing Laboratory at Chilton will not have to carry alone the burden of a national computing service.
Thereafter, says the report. 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.
AT the heart of St. Cross, one of Oxford University's two new graduate colleges, will be an electronic console linking the Fellows and research students with a giant computer 13 miles away.
The St. Cross console will be joined by landline to the Atlas computer at the Science Research Council's Laboratory at Chilton, near Abingdon. The link will make St. Cross, founded less than a year ago, the first centre in Europe deliberately to emphasise the non-numerical use of a computer as a research tool in arts subjects. The only other comparable centre in the West is the New York University Institute for Computer Research in the Humanities.
The project will be started while the only visible signs of the existence of the infant college are a body of 48 Fellows, a Victorian schoolroom, a large wooden hut and a newly-seeded lawn on the three-quarters of an acre St. Cross Rectory site.
The Master of St. Cross, Dr. W. E. van Heyningen, says that in the early days the console might be housed in another wooden hut, to be built possibly within a few months.
He said he had two chief considerations in suggesting the idea to the governing body, which enthusiastically discussed and unanimously approved it.
We don't want merely to ape existing colleges and - starting a new college in the latter half of the 20th century - we want to provide modern services.
Computers are infiltrating our culture in all directions. They are being used more and more for non-numerical application - for the interpretation of the Dead Sea Scrolls and the textual analysis of the Shakespeare First Folio, for work in law, medicine, psychology, management studies and other fields.
The idea began to grow on me that this was a service which could go a long way towards providing a soul for the college. It can act as a common denominator among studies in widely different fields.
The idea applies uniquely to a graduate college in Oxford because a college is a cross-section of the University, and graduates have research problems to which the computer might usefully be applied.
Dr. van Heyningen said Prof. Fox. head of the University Computing Laboratory, was enthusiastic about the idea of the difference between a university-based and a college-based computer.
He said: A University computing department is principally for dealing with numerical problems and is generally on the scientific side of the university.
It doesn't come naturally for people in the humanities to think about computers.
A graduate coming here will find that computers are in the air. He will find an atmosphere where it is natural to use them in research and perhaps he will decide to try to use one in his own subject - although, if he doesn't, we shan't throw him out.
St. Cross has held a seminar on the non-numerical use of computers. It was the first time I've heard a lawyer and a mathematician talking the same language, Dr. van Heyningen said.
I want to make it very clear that this is not a publicity gimmick and that St. Cross is not going to become a temple of the computer age.
We do not think they are capable of solving all problems, but they can help in the solution of more problems than people think.
We don't expect to be pioneers for long. We expect that other bodies will take up the same idea very quickly. The computer centre will take up 5,000 square feet of St. Cross's first permanent building, a tenth of its total area. In the end it will house more equipment than a single console linking to the Atlas at Chilton. There will probably be a number of smaller computers as well.
The college is to launch an initial appeal for between £lm. and £l½m. for this building, which will have halls, common rooms, seminar rooms, offices for the 50 Fellows, a Master's lodge and administrative offices.
After this, it will make a further appeal to build students accommodation in the part of St. Clement's designated by Oxford City Council for university residential development.
It looks like being the best year yet for the I.C.T.'s Atlas computers. With the acceptance by the UK Atomic Energy Authority of the Aldermaston Atlas, a period of some difficulty has come to an end and the Atlas is now beginning its work on secret data for the Atomic Weapons Research Establishment.
For about 10 months, the hardware at Aldermaston has been satisfactory. At the beginning of the year, both hardware and software went into test and these tests have proved successful. Acceptance has been back-dated to early January. Now the £1,600,000 computer has been handed over and largely paid for, the balance to be paid during 1966.
At the Chilton (Berkshire) establishment of the Science Research Council, a model Atlas installation - probably the largest working data processing system in Europe - is now on three-shift operating, doing 2,500 jobs a week spread over 20 hours daily and serving 25 universities as well as other bodies.
This £3,000,000 SRC installation was architect-designed to suit the Atlas with maximum job flow efficiency in mind. A final three months' testing period is now in progress and so far more than the 90 per cent availability demanded is being achieved.
At Manchester, the I.C.T.-owned prototype Atlas is used jointly by I.C.T. Computing Services Division and Manchester University. It handles between 1,700 and 2,000 jobs a week and brings in a tremendous amount of revenue for I.C.T.
One bureau Atlas user is the firm of Dorman Long who in just over two seconds do work which would take half an hour on their IBM 1401.
At Cambridge University, an Atlas II (the prototype of the Aldermaston model) was developed by the University and I.C.T., is maintained entirely by, and works solely for the University, with great success.
Unhappy stories about Atlas have centred almost entirely upon the London University Atlas. But the unfavourable comment, many people believe, is partly due to the attitude of the main users. They are not prepared to cope with a machine which, because of its complexity, is inevitably subject to a small number of breakdowns.
The London University machine handles around 3,000 jobs a week - two minutes being the average time spent on each. In the past three months it has been operating most of the time at more than 90 per cent average availability.
Experts agree that on such readings, Atlas makes better use of its central processor than any other computer. The IBM 7090 for instance averages 36 per cent efficiency whereas Atlas gives 64 per cent central processor efficiency.
An Atlas can compete, in power, with any computer in Europe. A simulation job can involve in the area of 12-billion program instructions, but occupy only 10 hours of Atlas time. And a Data and Control writer unearthing facts about Atlas discovered its running cost in electricity - just £1,000 a month!
NOT ONLY the British have difficulties with big computers. The science correspondence of the BBC programme Today made this point when he told the story of CDC's trouble with their giant 6600 at the European Centre for Nuclear Research near Geneva.
We hear that CDC have other worries too. Of eight 6600 systems in the field, it is reported that only three have been accepted. The others are therefore being used rent-free. Up-time for installation averages a little below 90 per cent and it is believed that this won't climb above 90 per cent before the end of the year.
Reason given is that complex systems like this take longer for new applications to work out and for component combinations to be fully tested. Same goes for Atlas.
