Jump Over Left Menu
- SERC's strategy in distributed interactive computing
- Developments in computing
- Symposium on computer graphics
SERC's strategy in distributed interactive computing
The emergence of cheap high powered single user computer systems with good interactive capabilities and standard communications interface heralds a completely new way for many research workers to achieve the major part of their computing requirements.
Within the next few years, many such systems will be available from different manufacturers. Consequently there is a danger of a variety of machines being employed in academic research, many of which are likely to have inadequate and incompatible systems software, leading to a dissipation of resources and considerable duplication of basic software development.
SERC sees a need for a coordinated plan to ensure that the academic community makes the best use of the resources available to it, especially its limited manpower. The Council has therefore decided on a policy of creating a common hardware and software base to act as a nucleus for future developments in single user workstation practice. Initially, the common software base will be Pascal and FORTRAN running under the Unix operating system implemented on the common hardware base of PERQ single user computers, linked locally by Cambridge Rings and nationally by the X-25 wide-area computer network. The PERQ will be joined by other machines as new products become available.
To assist in the implementation of this policy a small support team is being formed at the Rutherford Appleton Laboratory whose role will be to ensure that the systems software provided on the chosen machine types is satisfactory and to encourage the development of applications software through arrangements analogous to those of the, ICF's Special Interest Groups. SERC's Boards and committees will consider grant applications involving requests for single user computers in the normal way. If the scientific programme is approved, the application will be referred to the support team who will arrange the supply of a machine to the applicant. All purchases will come from Board or committee funds but the single point of procurement will enable bulk orders to be placed so that discounts and reductions in delivery times can be obtained. Central purchasing of PERQs, the UK manufacture of which is due to begin in early 1982, will enable ICL to increase the scale of its production with greater confidence. An integral feature of Council's overall strategy is that SERC and ICL will collaborate extensively on the development of future PERQ-type machines and on software systems relevant to the interests of the academic community.
The PERQ is an extremely powerful, single user workstation with a high precision display system which provides a significant improvement in the quality and speed of interaction. Although entirely self-contained, it also features standard communications interfaces offering network access to other local workstations or remote shared facilities. Its principal characteristics are:
High speed processor
Approximately 1 million 'high level' machine instructions per second giving around two-thirds the CPU power of a VAX 11/780. The CPU is micro-programmable for further speed gains.
High quality display
A4 size, 1024 x 768 pixel, high resolution black and white display featuring 60Hz non-interlaced refresh rate which enables pictures to be moved cleanly and rapidly as well as giving a significant improvement in the clarity of text and diagrams equal to a printed A4 page. User friendly I/O devices A 2-D tablet keyboard and voice synthesiser, allied to the high quality screen, enables a much improved man-machine interface to be created.
Large virtual memory
A 32-bit address, paged virtual memory system.
A 24 Mbyte Winchester disk and 1 Mbyte floppy give a single user a large amount of local storage capacity.
Local communication at 10 Mbits/sec via Cambridge Ring. Standard RS-232 serial and IEEE 488 parallel interfaces are also provided.
Common base policy
The whole academic community, not just computer science interests, is a major user and developer of software. The ease with which software can be developed and the extent to which it is easily interchangeable between machines can have a significant effect, therefore, on scientific productivity. The SERC believes that the programme it has initiated for single user machines will provide a way of removing many constraints on research output. Computing resources will be available locally with much greater freedom of access. Person to person and computer to computer links will be easier to establish and the prospects for collaboration and co-ordination improved. Commonality of hardware and software will maximise the opportunities for co-operation. A framework will be created in which software skills can be exploited, information on tools and techniques disseminated, and software made available in forms which can readily be adopted by the widest possible spectrum of users.
PERQ, and the new departure it represents, provides the incentive for creating a common hardware and software base in which the best of the existing tools, packages and techniques will be brought together in an overall framework, the effectiveness of which will be much greater than the sum of the individual elements. Work will be required in universities and at RAL to move existing software into the common base which will be complemented by selective purchasing and, in due course, by the direct results of research projects using the common base equipment.
The strategy which SERC is intending to pursue in distributed interactive computing is more than just a mechanism for standardising on one or two machine types. Its feasibility and timeliness are intimately related to a number of contemporary developments in computing.
The widespread availability of networks makes common access possible to special tools or facilities which can be provided only on a limited number of sites.
The pace of technological development means that over the next few years the cost of single user systems will diminish while their quality and capability increase. Today's PERQ is therefore seen as only the first machine in the common hardware base.
The compact PERQ single user computer at RAL
Developments in computing
The Council's Information Engineering Committee is responsible inter alia for research in the field of computing science. Described here are recent developments arising both from the Committee's specially promoted programme in Distributed Computing Systems and as further implementation of the recommendations of the 'Roberts Report'.
Distributed computing systems
The Specially Promoted Programme in Distributed Computing Systems (DCS), which started in 1977-78, now supports about 40 academic investigations, amounting to an investment of about Â£2.75M. The programme's objective is to gain an understanding of the principles of distributed computing and to establish engineering techniques to implement them. The programme is now entering its final phase - it will end in September 1984 - and effort is being concentrated on transferring the results of investigations to UK industry. To this end, the Council has appointed a part-time industrial coordinator with responsibility for establishing links between academic researchers within the programme and potential industrial users - principally in the computing hardware, software and applications sectors.
The DCS Industrial Coordinator is Mr. Fred Chambers, Logica Limited, 64 Newman Street, London W1A4SE. Tel: 01-6365440 ext 242.
A potential major product of the DCS Programme is the development of dataflow computing systems by Dr J Gurd and Dr I Watson of Manchester University who are world leaders in this field. Dataflow computers are able to execute many different parts of a program concurrently, achieving very high computing speeds. Potentially, they have much wider application than alternative parallel computers which rely on regular structure in hardware and software.
The Manchester Group has recently been awarded a four-year grant of Â£395k for the development of a prototype multi-ring dataflow computing system
Implementing the 'Roberts Report'
The Roberts Report (Proposed new initiatives in computing and computer applications SRC, 1979) recommended, in the national interest, that the Council should take action to stimulate industrially relevant academic research on the following topics within computer science:
- Software technology;
- Database utilisation;
- Systems reliability;
- Man-machine interaction.
Progress on implementing these recommendations has been considerable.
With the cost of computing power falling rapidly, there is an urgent need to find means of producing cheap, reliable software. The Engineering Board's Computing and Communications Subcommittee has appointed an officer to coordinate a research programme to respond to this challenge. The principal objectives of the programmes are:
- To stimulate more high-quality software engineering research;
- To improve the academic software technology base;
- To facilitate two-way technology transfer between industry and the academic world.
A major feature of the programme is the adoption of common hardware (ICL PERQ and Cambridge Ring) and software (Pascal and UNIX) bases. This approach will facilitate the transfer of results and ideas between research groups, reduce duplication of effort and enhance the attractiveness of the British products used.
The Software Technology Coordinator is Mr R Witty of the Council's Rutherford Appleton Laboratory, Chilton, Didcot, Oxon OX11 OQX, telephone: Abingdon (0235) 21900 ext 6218.
Following a community meeting in September 1980, to which academics and industrialists were invited, two major consortia emerged which have subsequently received awards totalling almost Â£500,000 for research on multi-site distributed data-bases.
One consortium, centred on East Anglia University, is concerned with a heterogeneous distributed database system. Its objective is to agree standards for logical data description and query representation at the internal system level. The other members of the consortium are based at the Universities of Aberdeen, Bristol, Cardiff and Kent.
The other consortium, centred on Aberdeen University, will establish a homogeneous distributed database. The aim is, at each site, to study different aspects (eg suitable query languages and storage mechanisms) of the same system. The other centres within the consortium are Heriot-Watt and Stirling Universities and the Polytechnics of Leeds and Northern Ireland.
A research group at Newcastle University, under Professor B Randell, has received long-term support from the Council for research on computer system reliability. The techniques developed are now at the stage where they should be transferred to industry. Consequently, the Council has awarded a grant of Â£148k to Dr T Anderson to design and construct, to full commercial standards, a command and control system embodying the reliability techniques developed at Newcastle. The investigation also receives support from the Ministry of Defence. It is hoped that this project will serve to demonstrate to industry the effectiveness and practicality of the reliability techniques developed with SERC support.
Man-machine interaction and knowledge-based systems
The Council is keen to receive proposals for industrially relevant research in man-machine interaction.
The possibility of establishing a coordinated research programme on knowledge-based systems, incorporating aspects of man-machine interaction, database utilisation, software technology and artificial intelligence, is being debated. Developments will be reported in subsequent issues of the Bulletin.
Further information on the programmes described above can be obtained from Mr J Monniot, SERC Central Office, Swindon, ext 2260.
Symposium on computer graphics
A symposium on computer graphics organised by the Computing Division of Rutherford Appleton Laboratory in October 1981 proved highly popular: 300 people applied for the 200 seats available. International experts were assembling at Abingdon for a meeting the following week of the ISO Graphics Working Group and RAL took the opportunity of their presence to hold the symposium. Five speakers gave presentations and took part in a lively panel discussion.
Peter Bono, from the Naval Underwater Systems Center in Connecticut, in a well-illustrated talk on 'Graphics Software - Past, Present and Future', discussed the division of available software into turnkey systems, application programs, graphics support and subroutine packages.
In contrast, Paul ten Hagen from the Mathematical Centre in Amsterdam, in 'Graphics Dialogue Programming', discussed the need for, and means of, separating the algorithmic and interactive parts of a program.
Another facet of computer graphics was revealed by Jim Michener of Intermetrics Inc. Cambridge, Mass in his talk on Graphics Real-time Applications Display Support. He showed the use of GRADS in a project for the US Navy, to design an advanced integrated display system for aircraft cockpits. Bob Hopgood's 'Road to Graphics Standards' traced the progress from the proposal in the early seventies that GINO-F should be a standard, to the current international efforts to refine GKS by resolving all the outstanding technical issues (a process that was finally concluded the following week).