EASE 90 has come and gone - all the comments made would suggest that it was successful; complaints - a few - we will get it right for EASE 91 at Birmingham. Make a date in your diary now - 25-27 March 1991.
Note the activity of the Software Distribution Service at Kent as important changes are planned - more details soon.
The latest update to the EASE Server Assessment gives certain important information; this is significant to anyone considering the purchase of these facilities.
1st Annual Symposium and Exhibition on Engineering Applications "Is the Future Parallel?"
It was a pleasure to attend the first EASE conference held at UMIST in a friendly atmosphere. The opening sessions given by speakers such as David May (Inmos Ltd), John Douce (University of Warwick), Georgio Molarini (Genoa University) and Mike Brady (University of Oxford) gave us a valuable overview of current directions in parallel processing, the future programme for EASE and developments in parallel algorithms for vision and robotics. It was particularly encouraging to hear that although little can be done to prevent Japanese domination of the present market, parallel processing may well allow the UK to regain the lead in future IT applications. Hardware developments, including the well publicised transputer, appear to be progressing well, but there is a need for more effort on software. Several speakers described the efforts of various funding committees to support engineering applications and we were surprised to hear John Loughhead (GEC and EMEC board of SERC) when he indicated there are not enough academic grant applications in the areas of plant analysis, simulation, design, control and monitoring!
On the other two days, the use of parallel sessions gave delegates some choice as to which tutorial sessions to attend, although I found it difficult to make that choice as they all seemed equally interesting. In the event, I settled for Real Time Control, Computer Integrated Manufacture and AI Techniques. Peter Fleming (Bangor) introduced a valuable session on Parallel Processing in Control, with contributions on SIMD and MIMD architectures, failure management, neural nets and AI. The CIM session was a lively affair, especially as Professor Wozny (Rennsselner Polytechnic) managed to present a vast number of excellent slides on concurrent engineering at the same time as packing to catch a flight to New York! I also enjoyed the tutorial session on AI which gave sound practical advice on how to approach knowledge-based system applications and select appropriate methodologies and software.
Alongside the conference, delegates had an opportunity to see some exciting new workstations and software demonstrated in the accompanying exhibition. This was much appreciated, though I doubt if the vendors would receive many orders form the poverty-stricken academics present!
Perhaps the most striking feature of this conference was the attendance approximately 260 participants, many of whom stayed for all three days. This is surely evidence of the major interest generated by EASE, by the topics chosen, and by the chosen speakers. The organisers are to be congratulated on their excellent organisation. I feel confident that the participants were well satisfied with the symposium and will look forward to the second one next year in Birmingham.
In spite of being organised at short notice EASE 90 attracted a total of 260 attendees who appeared to enjoy the experience.
In all fifty-eight speakers made presentations during the three days with 180 attending the combined session on the first day. The parallel sessions on the second day each attracted audiences of over forty, with congratulations to those hardy souls who scaled the heights to the eighth floor when the lifts refused to work.
The five tutorials also proved very popular, with a combined audience of 193. The most popular were the Introduction to X and Exploiting the Transputer, and it is intended that these will be re-run shortly on a regional basis.
EASE 90 proved that this type of event fulfills a need; EASE 91 at University of Birmingham on 25-27 March 1991 will be even better, make a note of it in your diary now.
On April 3rd, Sheffield City Polytechnic hosted an EASE Education and Awareness seminar entitled Computer Algebra and its Engineering Applications. The seminar proved very popular and was attended by more than 80 delegates from a wide range of backgrounds.
The seminar was chaired by Professor Derek Linkens (Department of Control Engineering, Sheffield University) who introduced the morning session by recounting his own experience with computer algebra systems.
Dr Chris Wooff (University of Liverpool Computer Laboratory) then gave a talk entitled, An Introduction to Computer Algebra Systems, in which he described the capabilities of the major computer algebra systems. These include basic algebra, calculus, equation-solving, matrix algebra and high-quality graphics.
Dr David Harper, U.K. Computer Algebra Support Officer, concluded the morning session with a talk which compared the five most widely-used computer algebra systems: REDUCE, Maple, MACSYMA, Mathematica and Derive. He outlined their origins and development, and described the features that distinguish each system from the others.
After an excellent buffet lunch, the seminar resumed. The afternoon session was devoted to examples of the application of computer algebra systems to engineering.
The first speaker was Dr Francis Wright (School of Mathematical Sciences at Queen Mary and Westfield College London) who illustrated the use of REDUCE to solve a problem in AC circuit theory and to manipulate a set of partial differential equations arising from a problem in fluid mechanics. Dr Wright showed how REDUCE could be used to verify a numerical solution of the first problem, and to detect a minor systematic error in a solution of the second problem obtained by hand.
The next speaker was MIle Christine Barbier, University of Durham Computer Centre, who illustrated the use of REDUCE to calculate the shape functions used in finite element analysis, together with their derivatives. MIle Barbier demonstrated how REDUCE's Fortran code generation package GENTRAN could then be used to convert these algebraic expressions into reliable and efficient Fortran programs.
Professor James Davenport and Mr Mike Dewar (School of Mathematics at University of Bath) then described their work on the IRENA project IRENA is a collaborative venture with the Numerical Algorithms Group (NAG) to enable the numerical analysis routines of the NAG library to be called from within REDUCE. When the project is completed, it will provide engineers and scientists with a tool combining the algebraic capabilities of REDUCE and the powerful numerical techniques of the NAG library.
The afternoon session was brought to a close by Dr David Harper who gave a brief description of the Computer Algebra Support Project which is based at the University of Liverpool. The remit of the Project is to provide advice and information on computer algebra to the UK academic community. Dr Harper said that the Project, now in its third year, had proved very successful and he hoped that the engineering community would take advantage of the services that it was able to offer.
In addition to the seminar sessions, there were demonstrations of Maple, REDUCE, Mathematica, MACSYMA, Derive and Symbolator throughout the day and these were very popular with all of the delegates. The SERC kindly provided a number of SUN workstations on which to demonstrate REDUCE and Maple.
The netlib system for distributing numerical software over Janet via electronic mail formally started on 1 January 1990. A full list of the available libraries may be obtained by sending a mail message to netlib@uk.ac.ukc whose body contains the single line
send index
The object of this service is to provide researchers with rapid access to specific numerical routines. For example, assume a user wishes to solve a system of linear equations, she consults his/her local numerical expert who advises that the Linpack routines dgeco and dgesl are appropriate. He?She then types the following:
mail netlib@uk.ac.ukc send dgeco dgesl from linpack
In a short time (hopefully less than an hour) mail comes back from netlib containing the double-precision Fortran subroutines dgeco and dgesl along with any other Linpack routines that these specified routines call.
There are a number of variants of the send command which control the amount of dependent software sent out by the system. There are also facilities available for retrieving the names of routines in the collection related to specified keywords and for retrieving electronic mail addresses.
Further details are given in the index me and in the article by Dongarra and Grosse in CACM Vol 30, 1987 pp. 403-407.
We are very keen to expand the software available and we welcome suggestions for further libraries/packages that researchers would find useful. We would also like to expand the directory of e-mail addresses. Both suggestions and addresses may be e-mailed to
netlib-suggest@uk.ac.ukc
| Requests | MBytes | |
|---|---|---|
| Dec 89 | 145 | 2.90 |
| Jan 90 | 606 | 12.61 |
| Feb 90 | 670 | 13.63 |
| Mar 90 | 775 | 17.31 |
| Apr 90 | 663 | 14.21 |
| Total | 2859 | 60.66 |
The table shows the usage in terms of the number of requests serviced by the system and the amount of software distributed on a monthly basis.
There have been 443 different users access the system of which about two-thirds have gone on to request additional information (ie more than just the main index file).
The total size of the database is currently 125MBytes but this will increase rapidly as new libraries (eg Toolpack, Applied Statistics, etc) are added to the system.
Since 1 January there have been a number of additions to the software database. A complete list of all changes to the available software may be obtained by sending the request:
send updates from updates
This file is regularly updated. We list below a number of the larger additions
In the near future we will be moving netlib to a SUN4. Having a stand alone system will allow us to provide the remote execution facilities which are currently available from the netlib system in the States. We are looking at ways in which this execution service (currently the US machine only offers a Fortran to C translation service) could be extended (eg a Fortran source formatter, a Fortran 77 Standard checker).
Issue 20 of this newsletter contained a Preliminary Report on the Server Assessment work which was undertaken during the last quarter of 1989. Because of problems in obtaining machines on loan, and also the lack of Coloured Book software, it was decided to extend the assessment into the first quarter of 1990. This article presents a short version of the report which was presented to the Computing Facilities Committee Technical Advisory Group (CFTAG) in April.
In the event, it was not possible to do much other than fill in some missing data, and test some of the other SUN servers. In particular, it was unfortunately not possible to obtain an example of the MIPS machine which we would have liked to test. In addition, those suppliers who had stated that their Coloured Book software would become available during the first quarter were not able to deliver. Thus it was still true that SUN were the only supplier able to meet the mandatory requirements. This has caused CFTAG to review those requirements (see below).
The final paper contained test data on the following set of machines:
| Supplier | Model | Processor | Memory | Hard disc |
|---|---|---|---|---|
| DEC 5400 | MIPS | 32Mb | 400Mb | |
| MIPS | M/120-5 | MIPS | 32Mb | 328Mb,663Mb |
| Solbourne | 5/801 | SPARC | 32Mb | 830Mb |
| Sun | 3/480 | 68030 | 32Mb | 892Mb |
| Sun | 4/370 | SPARC | 32Mb | 688Mb |
| Sun | 4/490 | SPARC | 32Mb | 2*1Gb |
The SUN4/370 was retested with a single SMD disc, rather than with multiple SCSI discs (as reported in the last article), since this was a more likely configuration for the server. In the event, the figures obtained were very similar, showing that the slower SCSI interface was offset by having more than one disc controller.
The SUN4/490 uses the new SUN IPI discs. These can take an appreciable amount of current at system start-up, and SUN supply the machine with 30A cabling. This requires extra electrical work at a site, and must be considered a disadvantage.
As stated previously, the preferred machine from MIPS would be the RC3240. The figures for the M/120-5 have been used in the following tables, and MIPS state that they expect the RC3240 to outperform the tested machine.
Benchmarks were run to test the performance of NFS under load. The write tests created a file and wrote 20Mbytes sequentially, using 8Kbyte buffers. The read test used this file to read data after performing a random seek. Clients were mostly SUN3s, but for tests involving more than seven clients, the more powerful SUN4 was added. One of the existing RAL servers (SUN4/280, 16 Mbytes, no longer supplied by SUN) was used as the reference machine.
The key to the tables is:
| Code | Range of ratios to SUN4/280 |
|---|---|
| Blank | 0.9-1.2 |
| + | 1.2-1.5 |
| ++ | 1.5-2.0 |
| +++ | 2.0-3.0 |
| ++++ | > 3 |
| - | 0.7-0.9 |
| -- | 0.5-0,7 |
| * | No result |
| Supplier | Model | Number of Clients Served | |||||
|---|---|---|---|---|---|---|---|
| 1 | 2 | 4 | 6 | 8 | 10 | ||
| DEC | 5400 | -- | -- | -- | -- | -- | * |
| MIPS | M/120-5 | - | - | - | |||
| Solbourne | 5/801 | ++ | ++ | - | - | ||
| Sun | 3/480 | ++ | ++ | + | |||
| Sun | 4/370 | ++ | +++ | - | - | ||
| Sun | 4/490 | ++ | +++ | +++ | +++ | +++ | ++ |
DEC was unable to complete the test with ten clients due to a shortage of disc space. Both the Solbourne and the SUN4/370 showed a considerable drop in performance between the 2-client and 4-client tests (no reason for this has yet been established). This behaviour was not followed by the reference SUN4/280, however (which uses an earlier version of the SPARC chip). It was also interesting to note that the reference machine performed better than most under heavy load, even though it had half the memory of the other contenders. The results of the write tests are given below:
| Supplier | Model | Number of Clients Served | |||||
|---|---|---|---|---|---|---|---|
| 1 | 2 | 4 | 6 | 8 | 10 | ||
| DEC | 5400 | - | -- | -- | - | - | * |
| MIPS | M/120-5 | +++ | ++++ | ++++ | ++++ | ++++ | ++++ |
| Solbourne | 5/801 | + | + | + | + | ||
| Sun | 3/480 | + | + | + | + | ||
| Sun | 4/370 | + | + | + | + | ||
| Sun | 4/490 | + | ++ | +++ | ++ | ++ | ++ |
Again, DEC was unable to complete the test with ten clients. The change in performance under load is nothing like as marked as in the previous case.
The exceptional results produced by MIPS have yet to be explained. It is most unusual to have write figures which are better than those for read. It may be significant that MIPS is the only System V based machine, running the new fast file system (ffs). Further tests suggest that data may not in fact have reached the disc when the completion signal is received by the client, which will obviously affect the timing.
The following table presents a summary of the general characteristics and performance of each machine (the more asterisks the better). A - indicates an apparent lack of the feature. The cost information for the SUN4/280 is based on old data, since the machine is not currently supplied. The assessments are non-competitive, and do not represent a ranking.
| Machine | Basic System | Network | Perforance | Cost | |||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| 1 | 2 | 3 | 4 | 5 | 6 | 7 | 8 | 9 | 10 | 11 | 12 | ||
| Sun 4/280 | *** | *** | **** | **** | **** | **** | *** | *** | **** | *** | *** | *** | (*****) |
| DEC 5400 | **** | **** | **** | **** | ** | ** | *** | ** | ** | ** | *** | - | **** |
| MIPS M/120-5 | ***** | *** | *** | *** | *** | *** | - | *** | *** | **** | - | * | - |
| Solbourne 5/801 | ***** | **** | *** | **** | ** | *** | - | **** | *** | **** | *** | - | ** |
| Sun 3/480 | ** | *** | *** | **** | **** | **** | *** | **** | **** | **** | *** | *** | *** |
| Sun 4/370 | **** | *** | *** | **** | **** | **** | *** | **** | *** | **** | *** | *** | **** |
| Sun 4/490 | ***** | *** | *** | **** | **** | **** | *** | **** | ***** | ***** | *** | *** | ** |
| Code | Title |
|---|---|
| 1 | Hardware |
| 2 | Dump/Archive Facilities |
| 3 | Administration |
| 4 | NFS/Yellow Pages |
| 5 | LAN |
| 6 | WAN |
| 7 | IP Routing |
| 8 | NFS Read (2 clients) |
| 9 | NFS Read (4 clients) |
| 10 | NFS Write (2 clients) |
| 11 | Yellow Pages Performance |
| 12 | Coloured Books Functionality |
CFTAG discussed the report at the April meeting, and have made the following recommendations:
As in previous assessments, a more detailed version of the report will be published.
The first meeting of the CFD Community Club was held at RAL on Thursday March 22. It was chaired by Professor G Hammond (Bath) and was attended by more than 60 people. The meeting consisted of presentations on support or use of CFD in those committees of Engineering Board which fund projects in CFD, followed by a discussion on some of the technical issues to be addressed by the CFD Community Club.
Professor Bradley (Leeds) gave an overview of the current state of development and use of CFD in the area of interest to the Electromechanical Engineering Committee. He mentioned some examples of problems which can now be tackled, eg coupling of unsteady fluid flow and buckling of structures, use of adaptive finite element techniques in aerodynamical flow, and especially 3-D time dependent flows in a reciprocating engine, including modelling of combustion and the stressing due to knock shock waves. He highlighted developments in experimental verification of simulations by explaining the use of laser-induced fluorescent sheets to provide measurements in the study of flames, and commented that it is now possible to observe computational effects first, and to validate these later using experimental measurements.
Professor Bradley mentioned that six parallel computing systems are currently being purchased through Daresbury Laboratory. A LINK Programme in CFD is being formulated, and a CFD Coordinator and Monitor will be appointed soon. The Coordinator will be mostly involved with the computational aspects of CFD, and is expected to be based at RAL. There will be improved liaison with other initiatives in CFD, including ERCOFT AC (the European Research Community on Flow Turbulence and Combustion).
Dr Knight (Birmingham) surveyed the practical use of Computational Hydraulics and CFD in a number of engineering topics, viz Rivers, Estuaries, In Shore and Off Shore Phenomena, Waves, Lakes, Groundwater, Irrigation, Buildings, and Others (including pipe networks and urban drainage). He gave some examples, ranging from increased rice production in the Philippines to improved coastal protection and recent flooding in the Thames Valley. He presented the modelling strategies used in Computational Hydraulics and applied them to the modelling of estuaries and waves, where the problems have very complex boundary conditions. He explained that SERC is now mid-way through an eight-year programme in which a major experimental facility has been set up at Wallingford.
Dr Weatherill (Swansea) then outlined the background leading to the setting up of ERCOFTAC in June 1988, and described briefly its aims and structure. It emerged from the conviction that for European industry to be competitive there needs to be more cooperation and in particular a strong interaction between the research community and industry. There are now twelve pilot centres, including two in the UK. The UK Centres have been re-organised into UK North, coordinated by UMIST, and UK South, coordinated by RAL. There are now 52 research groups and 23 industrial companies affiliated to ERCOFTAC in Europe, with 13 research groups and four companies in the UK. There is a small subscription fee to join ERCOFTAC, which provides its only income. Several Special Interest Groups have been set up to suggest possible research topics.
Dr Davies (British Gas) gave some examples of the use of CFD at British Gas. These included Modelling of Gas Dispersion, Process Development, Natural Gas Utilisation, Burner Design, etc. He gave details of comparisons of experimental data with predicted results from commercial packages and in-house software. British Gas has recently taken delivery of an Alliant 4 processor system and a Silicon Graphics workstation. At present they feel that it is cheaper to add more hardware than to optimise software for a particular hardware system.
The discussion session focussed on the following topics - Visualisation, Integration of Software, Parallel Processing, Awareness and Support, and Numerical Accuracy. There was a general recognition of the need for portability of visualisation systems and of the use of graphics standards where possible. There was interest in using the flow visualisation software developed by NASA Ames Laboratory, which currently runs on the Silicon Graphics IRIS Workstation coupled to the CRAY 2 and CRAY Y/MP. It was recognised that porting to other hardware may be a problem in the future. The meeting agreed to hold a Workshop on Visualisation in CFD. Details will be in the next issue of this Newsletter.
The meeting agreed that RAL should acquire some of the existing commercial codes and the public domain system TEACH (including source code), and evaluate their strengths and weaknesses, and that the Advisory Group should attempt to specify the needs of the community and look at the attributes of the type of code which would meet those needs. There was general agreement that SERC should focus on providing a good parallel software environment to ensure that software should run efficiently on a variety of parallel systems. At least six people present are currently developing parallel versions of CFD codes.
Finally the problem of separating out the errors caused by the numerical methods from those caused by use of the current turbulence model was discussed. It was agreed to hold a second workshop on Numerical Accuracy later in the year.
As part of the AI Support for Engineers project, the Artificial Intelligence Applications Institute, University of Edinburgh is producing a series of reports on the evaluation of AI tools. As announced in Issue 2 of the AI For Engineers bulletin, two new reports have just been added to this series. One report is on ART-IM and the other one is on Xi Plus. These and a previous report on Crystal, and the survey on PC expert system shells' external interfaces are available free of charge on request. Therefore to receive a copy please contact us at the address given below.
ART -IM (Automated Reasoning Tool for Information Management) is a powerful program development environment for the creation of knowledge-based systems. Produced by Inference Corporation, it offers a range of AI programming styles in an integrated development and debugging environment. Essentially it is a rule-based programming tool, supporting an extensive patternmatching language which is invoked by patterns held within the rules. These rules become active whenever the patterns are matched by information held within the ART-IM memory, causing procedural code to be executed. ART-IM provides an easy-to-use procedural language together with a wide variety of pre-defined functions. Additional functions can be defined using ART-IM procedural language or the C programming language. Object-oriented programming is also available.
ART-IM has been designed specifically to support knowledge based systems development and delivery on PCs and mainframes, conforming to IBM System Applications Architecture standards. There is strong emphasis placed on its ability to be embedded within large existing mainframe applications. It requires a PC with 640K of main memory plus 1 to 2MB extended memory for the development environment and a hard disk with at least 8MB free space. Applications can be delivered on a basic PC as executable C code.
This product has been developed from ART which is a well-established hybrid toolkit first released in March 1985 and targeting mainly at R & D groups developing large knowledge bases. The report will describe some of the features of the ART-1M language, the development and the delivery environment, and the quality of the documentation.
Expertech's first product, known as Xi, was first released in 1985, and was written in Prolog. This was superseded by Xi Plus, the current version of which is written in C. Again, Xi Plus is a rule based tool but because it is a smaller system than ART-1M, it requires less computing resources and it runs on a basic PC with 512K of main memory with a hard disk. Xi Plus supports the idea that an application may be made up from many different smaller knowledge bases.
Over the years it has acquired a large user-base. The application synopses provided by the vendor suggest that it is used extensively in the banking, finance and government sectors, although engineering and manufacturing applications are also included. This report looks at the facilities provided in Xi Plus from the knowledge engineer's point of view.
Each course will be held in the Department of Computer Science, from 9.30 to 17.30 each day. All practical work for the courses is carried out on Sun SPARC stations in the new PEVE training room.
The cost includes a full set of course notes, expert tuition for both practical and theoretical sessions plus lunch and refreshments on each day.
In the article on the closure of the Interactive Computing Facility in the last newsletter, a series of articles which will summarise the ICF was promised. This is the first article, issued as a separate insert for convenient collation with future articles.
The Engineering Board set up a Working Party in 1973 to establish the computing needs of engineers in the universities and polytechnics and to see if a rational framework might be constituted for the provision of facilities to meet these needs.
The essential findings of the Working Group were that there was a pervasive requirement for interactive computing in engineering research, that the importance of the requirement was inadequately recognised in the facilities available to universities and polytechnics and that the provision of facilities on a piece-meal basis through individual grants was not cost-effective. The Working Group recommended inter alia that the SERC should set up a centre, based at an SERC establishment, to provide facilities for interactive computing to engineers in universities and polytechnics. The centre would moreover provide the stimulus for the promotion and development of research activities over a wide range of engineering disciplines where interactive computing has a powerful role to play.
It was recommended that a small Technical Group should be set up to draw up the detailed proposals for such a centre.
The terms of reference of the Technical Group were:
The membership of the Group was:
Technical support was provided by:
The Technical Group set up a Software Sub-Group to advise it under item 6. of the terms of reference. The membership of the Sub-Group was:
The Group reported in 1975 and concluded that there was a substantial and immediate need for additional interactive computing for engineers engaged on SERC funded research. The potential user population was seen as 900 in 1975, rising to 1300 in 1980. The Report described a plan by which SERC could provide an integrated and coordinated service for interactive computing (the Interactive Computing Facility - ICF).
A summary of the main recommendations follows:
To meet the computing needs for 1976-78:
To anticipate developments beyond 1978:
To satisfy other requirements:
It was foreseen that by 1978, the most cost effective means of providing interactive computing would be through multi-user mini computers interconnected by a network. Multi-user minis were new (Dec had not yet announced the VAX range of machines). Existing facilities available to the academic research community offering minimal interactive computing capability were small, single-user PDP11s.
It was recommended that a review be made in 1978 to establish definite plans for the period after this.
The next article in this series will cover the Multi-User Mini Assessment and the Implementation of Phase I of the ICF (1976-1978).
This leaflet explains how DTI and SERC organise their support for information technology.
The Joint Framework for Information Technology (JFIT) was established by the Department of Trade and Industry (DTI) and the Science and Engineering Research Council (SERC) in 1988 to extend coordination of DTI and SERC policies and support for Information Technology (IT). It aims to create coherent policies, objectives and priorities for IT research activities in universities, polytechnics, companies and national research institutions.
A key element in the coordination of policies and programmes is the joint committee structure which advises both DTI and SERe. This is headed by the Information Technology Advisory Board (ITAB) which provides advice on strategic policy issues, research priorities and resource allocation. ITAB delegates responsibility for specific policies and programmes to committees on Devices, Systems Architecture, Systems Engineering, Control & Instrumentation and Education & Training. These committees in turn work through more specialised sub-committees which are responsible for the detailed supervision of specific programmes and the approval of individual projects. Each committee consists of about a dozen members drawn from industry (including both users and suppliers of IT) and the academic sector; most committees maintain formal liaison with the Ministry of Defence.
The following activities in Information Technology are currently being supported under the Joint Framework:
In 1989/90 these activities will be supported by about £80M split roughly 60% for research and 20% each for technology transfer and education and training.
ITAB has determined three priorities for JFIT:
There are two types of research programmes supported under JFIT through DTI and SERC: collaborative programmes involving industry and the academic sector and non-collaborative programmes funded solely by SERC for projects at higher education institutions.
LINK aims to stimulate collaboration between industrial and science-base partners on innovative research projects in key areas of technology. Government funding may not exceed 50% of programme costs. Currently there are 5 LINK IT R & D programmes. In each case proposals are invited on topics set out in a predetermined workplan. The following LINK programmes are open for applications for the periods indicated:
(Funding: DTI £9M, SERC £3M, Industry £12M)
Launched in February 1988, this 5 year programme aims to improve the performance of microelectronic and optoelectronic devices and components by making better semiconductor materials and fabrication methods available. SERC also supports a complementary programme of fundamental research in Low Dimensional Structures and Devices.
(Funding: DTI and SERC each £5M, industry £1 OM)
Launched in February 1988 for 5 years, its aim is to support fundamental research into the organisational and functional properties of molecular systems and the development of suitable molecular and macromolecular materials in electronics and related areas such as optoelectronics.
(Funding DTI H5M, SERC £3M, industry £6.5M)
Launched in June 1988 for 3 years, this programme is prompted by the convergence of telecommunications and radio technologies which is generating a new industry of personal communications.
(Funding DTI & SERC £5.5M each, industry £11 M)
Launched in April 1988, this 6 year programme is concentrating on research and development of instrumentation and measurement technologies where there is an emphasis on system integration.
(Funding: DTI £1 OM, SERC £5M, industry £15M)
Announced in April 1989, this 5 year programme is building on the optoelectronics science base provided by JOERS concentrating in particular upon optical communications systems and subsystems, and optical information processing.
These programmes are designed to encourage pre-competitive collaborative research in new technologies and to accelerate their application. Government funding may not exceed 50% of programme costs. Currently there are 3 Advanced Technology Programmes being supported under JFIT:
(Funding: DTI £13M - still open for applications) Launched in January 1987, this 3 year programme supports collaborative projects from industry and other organisations across the whole range of electronic applications of GaAs. Contact: Dr M I Ridge, *01-2152750 High Temperature Superconductivity Advanced Technology Programme (Funding: DTI up to £8M - still open for applications)
Launched in January 1988 for 3 years, this programme supports research into novel high temperature superconducting materials.The programme is a major part of the UK effort in this area. It is coordinated with the work of the National Committee for Superconductivity which is concerned with bridging the gap between industry and academe.
(Funding DTI £34M, SERC £27M - allocated on a basis of specific calls for proposals)
Launched in June 1988 and scheduled to run for 4 years, this programme concentrates on selected topics where research is deemed to contribute significantly to industrial competitiveness. It aims to increase the number of smaller firms engaged in IT research, strengthen links between industry and the academic sector and increase the participation of users of IT in IT research. There have been two calls so far and a third in 1990 is likely. Topic areas are as follows:
National Electronics Research Initiatives are collaborative precompetitive research activities in fundamental areas of science and technology. Work is carried out at one location to take advantage of specific facilities and collective expertise. By bringing together researchers seconded from companies the UK has the opportunity to pursue new research more effectively in world terms. Government funding may not exceed 50% of programme costs. DTI is currently supporting the following initiatives:
This £350M pre-competitive collaborative programme started in 1983 and is now in its final stages. The Government element of the funding is provided by OTI, SERC and MoD. The programme is no longer open for applications.
The DTI/SERC funded Joint Optoelectronics Research Scheme (JOERS) is also now in its later stages. Long term precompetitive collaborative research projects have included work on novel materials and components for advanced fibre optic systems, integrated optics, optical signal processing, optoelectronic integrated circuits, sensors and displays. The scheme is no longer open for applications.
SERC's objective of sustaining standards of research in HEIs is traditionally achieved through the support of projects which do not require collaboration with industry. This continues within JFIT. Some support is for fundamental research; some for more strategic work. Grants are made either in response to unsolicited proposals, or as part of a series of co-ordinated programmes. In the responsive mode, SERC is, in principle, open to proposals in any technical area. For purposes of illustration, the range of such support includes the following topics, but is not limited to them:
Silicon technology, compound semiconductor technology, optoelectronics, microelectronics design, novel devices, advanced processing technology.
Artificial intelligence, cognitive science, natural language, logic, formal methods, software engineering, reliability, dependability and metrics.
Parallel and novel architecture, neural networks, speech, vision, communications, signal processing, distributed systems, multimedia systems, switching technology, networks.
Control theory, control system design, principles of measurement.
This is a substantial programme concerned with understanding quantum mechanical effects associated with low dimensional structures and their application to electronic devices. It is supported jointly through JFIT and SERC's Materials Science and Engineering Commission.
Since 1978 SERC has maintained microelectronics facilities to support the Council's research programme by providing access for the academic research community to advanced microelectronics design and fabrication. The facilities are distributed amongst selected university departments and at Rutherford Appleton Laboratory.
This programme began in 1988 and is expected to run for five years. It aims to stimulate research in high speed, high performance multiservice and multimedia communications networks.
This initiative is designed to stimulate research into control systems design, mainly through the development of a common software environment within which a variety of design tools may be embedded and which will facilitate the process of technology transfer to users.
This initiative was established to encourage understanding between logicians and computer scientists, with the objective that IT might be put on a more vigorous theoretical foundation.
Formally, this initiative lies outside JFIT, but is strongly related to it. Launched in 1989, three research councils - SERC, MRC and ESRC - have combined to encourage a multidisciplinary approach to research in cognitive science and human computer interaction.
ITAB and the individual technical advisory committees seek to coordinate the national programme with other international initiatives. The major European programme is ESPRIT, although information technology also appears in the RACE programme (Research into Advanced Communications in Europe), BRITE (Basic Research in Industrial Technologies for Europe) and EURAM (European Research in Advanced Materials). It is envisaged that all these European Community programmes will be extended in the new Framework programme (1990 - 94) and so coordination with the national programme will continue to be an important activity.
Another major international activity is the Joint European Submicron Silicon Initiative (JESSI), a £25bn, eight year EUREKA programme aimed at developing a wide range of submicron microelectronics products. Once again this is being closely coordinated with the national programme.
There are two major technology transfer programmes targeted at users of IT, the Advanced IT Technology Transfer (AIP) programme and the Open Systems Technology Transfer (OSTT) programme. Both were launched in the summer of 1988 to run for 3 - 4 years, with a budget of just over £ 12M each. There is also a smaller technology transfer programme which relates to the academic sector only.
Within the academic community SERC encourages the use of advanced IT in other areas of scientific research through collaborative programmes between IT and other research disciplines.
There are three main areas of education and training which fall within JFIT. The first is the traditional support for post graduate studentships and courses provided by SERe. An annual budget of £15M covers a variety of research and training studentships such as CASE (Co-operative Awards in Science and Engineering) and one year IT conversion courses designed to attract good quality graduates from other disciplines.
Secondly, DTI tries to ensure that other public support for education and training reflects industrial priorities. It also encourages firms to place a higher value on training and retraining. More specifically, DTI supports a number of Regional Electronics Centres and operates a limited number of specific training programmes such as that in Electronic CAD Software (ECAD).
Finally, the DTI and SERC jointly operate the Teaching Company Scheme (TCS) which, although not itself within JFIT, is developing a specific interest in IT.
Eligibility rules vary between programmes, but generally SERC funds are available for UK HEIs, and DTI may support any organisation capable of carrying out the research and exploiting the results in the UK. Any organisation, public or private, may also participate on an unsupported basis.
