This longer, 12-page issue is the result of many interesting and diverse contributions received recently. With more equipment assessments, an EMR contract on offer, a community club report and many training opportunities, this issue must surely have something for everyone. If your particular area of interest is not usually covered in the Newsletter, please let me know.
If you have not seen this Newsletter before and wish to get on the mailing list, please send your name and address, preferably bye-mail, to me.
This article gives an overview of a Bristol Polytechnic Teaching Company Project which seeks to create design tools to assist in the design for manufacture of civil aircraft. The project can loosely be described as market led research.
Present emphasis on designing for manufacture originates from the designer's historical separation from commercial matters. In today's harsher economic environment competitiveness in product pricing dictates a radical overall of the design processes and substantial reeducation of designers. The design stage relies on the use of people as creative problem solvers, for whom tools must be provided that steer thinking towards materials and geometry which ease the burden of production without compromising function.
The project is split into two distinct parts, firstly the collection and assimilation of manufacturing expertise and secondly the representation of the expertise in the tool. Experimentation with a PC based expert system shell rapidly showed that the task of translating design for manufacture knowledge into software would not be a small task. The number of different manufacturing processes found, coupled with their depth, was seen to pose a formidable knowledge and data handling problem.
Psychologists have devised a number of appealing techniques which can be applied to the engineering domain, eg the card sort, multidimensional scaling and the reparatory grid. Essentially they allow structuring of conscious but automatic thought processes. Manufacturing engineers are not always the most proficient communicators, but when asked to sort Q-cards into order and to reason about that order then any barriers are overcome. The outcome can be used to focus attention on the key aspects of the expert's domain.
It has been found that it is the relationship between a part's exhibited features and the processes used to form and assemble them that is critical when manufacturing. The component parts and production processes used in the aircraft industry lend themselves to an object-oriented representation. Following an evaluation of a range of software packages the knowledge engineering tool KEE was chosen as the host environment. In practice, KEE has shown itself to be the ideal choice.
The importance of the way in which the designer interacts with the system cannot be understated, since acceptance of design for manufacture will be largely dependent on the ease with which the designer can operate the system. Linking the DFM advisor system directly to the CAD system resolves the problem by reducing the designer's involvement in the data flow. The solid modeller Pro/Engineer was chosen as the front-end since its easy to use objected oriented approach captures designer intent and minimises the need for feature recognition.
The DFM system user interface is maintained as a third UNIX-X11 process and presents graphical and textual displays. This module was written with the aid of Sun's X-view toolkit and developer's guide code generator.
Data exchange between the processes conforms to a defined protocol and consists of three layers managing socket communications between the UNIX processes, geometric data packet format and command action requests. Throughout the project heavy emphasis has been placed on the generation of reusable code largely because the greater care taken in its conception contributes to higher quality and more robust operation.
Early this year a careful examination of hardware revealed only a Sun 3 to be capable of fully supporting the link. The difficulty stems from the radically different nature of LISP and C, the base languages of KEE and Pro/Engineer. The project software will be ported to SPARC in August.
The industrial partner is extremely happy with the results to date and system trials will begin early in 1991. In conclusion it may be said that the project has uncovered many areas of manufacturing where knowledge based technologies could have revolutionary influence.
The Computing Facilities Committee (CFC) has provided funding to produce up to four Transputer Demonstrators in Real Time Control Engineering over the next two years. The funding is to cover a total of two-man years of research effort (ie six-man months per project) plus recurrent costs.
One such project has already been funded to produce a hexapodal robot demonstrator (Prof J 0 Gray, University of Salford). On the advice of the Control Transputer Applications Community Club (CTACC), CFC has decided to invite further proposals for a six-month EMR contract aimed at producing a demonstrator, which is not in the Robotics area, to show the potential of Transputers for Real Time Control Engineering Applications. Ideally the demonstrator will cover a range of technical domains and exhibit clearly the need for concurrency based on a Transputer architecture. It should show that the use of multiple Transputers has allowed the realisation of a novel and stimulating Real Time Control Application and has led to a significant improvement in the performance of an existing application. The demonstrator should be self-contained, transportable and have visual appeal. Proposals for desk-top applications are particularly welcomed.
It is likely that a contract supporting one RA full time for six months (or equivalent) would be used to develop an existing piece of equipment to fulfill this role. Similarly, it is likely that the demonstrator will be based on existing research. It is anticipated that the successful bid will not exceed £15K, including overheads. An award would be conditional on the deliverable being made available for two years beyond the end of the contract.
An outline proposal is invited, not exceeding two sides of A4 paper, and should include a schedule of costs, timescales and deliverables. Proposals should be sent to me by 22 August 1990 and will be considered by the Computing Facilities Technical Advisory Group at its meeting on 26 September 1990, with advice from the CTACC Executive.
I read the article in Issue 23 on the EASE Server assessment with interest.
To use the phrase more normally associated with the advertising of books, for example, Coloured Books are of course available from all leading suppliers. In particular, I think that DEC and MIPS do offer Coloured Books for the equipment surveyed, although I understand that this was not apparently the case when the survey was originally done. I suggest that those wishing for more information should use their normal contacts with the suppliers concerned.
As Phil correctly points out, both DEC and MIPS now have the Coloured Books software, as does Solbourne, the third supplier that took part in the exercise without the software at that time. This was not the case when the Report was presented to CFTAG in April. Since the software has been written for many suppliers by the same external source, it is likely that others, who originally declined to take part, could now do so in a repeat of the exercise planned for the end of 1990. The full report is now available from the RAL Library as RAL-90-032.
CHEST has concluded several important deals in recent months, and the following list outlines the new agreements:
Information about these agreements is now being mailed to Computer Centres and CHEST Site Contacts at all institutions, to whom initial enquiries should be made.
The CHEST Directory will shortly be mounted on the NISSPAC service, providing the user with a much more sophisticated interface to information on CHEST deals than is currently available on the NISSBB. We are also planning to provide a version of the Directory on the NPDSA Service at Lancaster which can be easily downloaded. The current 1989190 printed edition of the CHEST Directory will be superseded this Autumn by an updated volume, and the 1989190 version is now available to members of the Higher Education Community at the cost of postage and packing only. A bargain not to be missed at £2.00 - contact Deborah Sawyer in the CHEST Office at Bath (D.H.Sawyer@swurcc) for copies.
NISSPAC evolved from the original NISS Software and Dataset Catalogue, and is a more advanced system providing a common, easy to use, interface to a range of related databases. We also plan to provide an additional facility which allows the user to interrogate all the databases simultaneously. NISSPAC currently includes three datasets:
If you have a connection to JANET and you wish to use NISSPAC, the command call niss.pac (numeric address 000050300009)) will access the host machine. A single dot prompt indicates a successful connection, and typing L NISSCAT will enter the opening NISSPAC menu.
As a consequence of the continuing popularity and high levels of usage of the NISSBB, an upgrade to its underlying hardware has become necessary. Thanks to further support from DEC, we have been able to purchase a Micro VAX 3800 as a replacement for the existing Micro VAX 2. The service will be switched to the new machine during the summer months.
As part of a drive to make access to our online services as easy as possible, we plan to launch the NISS Gateway by midsummer. Once the Gateway is fully implemented, users will need enter only a single command, call niss, at the PAD in order to gain access to all NISS's online services. The Gateway will provide a top-level menu from which a selection (eg NISSBB, or NISSPAC) can be made, with ensuing automatic connection to the preferred service. The call niss (at numeric address 00006220000) command can continue to be used to access the NISSBB until the Gateway is implemented. Following that, the NISSBB will then have a new NRS name, UK.AC.NISS.BB, which can be used if direct access to the Bulletin Board is required.
Mr D J Johnston and Mr B Henderson (RAL)
Attendees should have some knowledge of programming for conventional serial machines, with awareness of parallel processing being helpful. Topics include an overview of the transputer, characteristics, types, etc. Examples of methods of exploitation will be given.
Dr P Chung and Dr T Lydiard (AIAI, Edinburgh)
This tutorial will introduce participants to various practical aspects of building knowledge based systems. Topics covered will include how to acquire and analyse knowledge, how to represent knowledge, how to select a development tool, how to create a knowledge base and how to integrate a knowledge based system with existing systems.
Dr K Robinson (RAL)
This tutorial aims to give the non-expert an overview of the X-window system, its capabilities and drawbacks, and where to go for further information. Some exposure to windowing systems will be assumed, though not necessarily to have programmed in such an environment.
This free seminar will present an introduction to Object Oriented Programming (OOP), a comparison of 00 languages and environments, and four presentations covering different aspects of 00 techniques relevant to engineers. Fuller details of the programme and registration form are enclosed with this issue of the Newsletter.
This programming technique is becoming increasingly popular; it has been extensively used in the area of user interface design in particular (eg in the Apple Macintosh user interface). It is, however, generally applicable to all areas of software development. Briefly:
Thus code generated using the object oriented approach is:
A number of object-oriented programming systems (OOPS) are available, such as Smalltalk, C++, Eiffel, Objective-C, etc, on equipment as varied as PCs, workstations and mainframes.
A Research Assistant is required to work on a project funded by SERC to develop advanced visualisation techniques for computational fluid dynamics. The post, in the Department of Civil Engineering, University College of Swansea, is for two years from October 1990 and will be carried out under the supervision of Dr N P Weatherill. The work will be undertaken in close collaboration with the Informatics Department, Rutherford Appleton Laboratory. The work will involve the use of advanced graphics techniques and hence prospective candidates should ideally have some experience in this area.
One year ago the Computers in Teaching Initiative (CTI) Centre for Engineering began to collect information about the use of computers in under-graduate and taught postgraduate courses in university departments. Statistics were compiled about the usage of software in different categories. Several reports have appeared in this newsletter but nothing has been written about the nature of the software. Departmental contacts were invited to list items of software but were not required to provide a detailed inventory; nonetheless 991 reports of software were received. These reports boil down to over 500 different packages and of these, close to 200 were indicated to have been developed in-house. The top ten packages, by number of reports are illustrated in figure 1.
All of these packages are commercial and several are quite old. Other popular packages with 5-7 reports include: AsEasyAs, AutoSketch, Chi Writer, FirstWord Plus, GHOST, Integrated 7, Latex, LUSAS, MS Word, MS Works, NAG PC50, Process, Silvar Lisco, Simbol, SPICE, SPSS, SuperCalc, WordPerfect, and View.
In contrast to the commercial packages, software packages that have been written by departments tend to remain within that department or were simply not reported. It is likely that few engineers pay much attention to the birthplace of the packages they use everyday. Specialty packages that are used during discrete periods of the year to fit in with the curriculum may rest idle and forgotten on computer discs. Three departments in different universities reported using a non-commercial finite element package called PINEL. This is such an obvious name for an FE program that it probably refers to 3 separate programs. A linear programming package has been reported in the same way. A word-processing package called SCRIBE is used by two departments at Edinburgh University. CTI Project software developed at more than a dozen sites around the country has been made available to other departments as a condition of the original grants, yet not all of it was reported. Similarly, Engineering Science Program Exchange (ESPE) software has been dispatched from Queen Mary and Westfield College to fifty institutions around the country and is known to be in regular use.
Given the amount of effort that goes into developing a robust user-friendly pedagogically sound teaching package, it is surely in everyone's interest not to duplicate such effort. For this reason the Centre software catalogue highlighted the software that has been written within university departments to meet a specific teaching need but packages have only been included if they are known to be available to other departments.
Interesting questions are how many packages and of what sort are bought in and where departments are writing their own. The answers are illustrated in figures 2 and 3. Each discipline section of a bar is composed of n number of distinct and separate packages. The numbers in brackets after the disciplines in the legend are the number of departments in that discipline who returned completed Questionnaires to the Centre.
As one might expect, where suitable commercial software exists, it is bought in: hardly anyone writes their own spreadsheets, word-processing software or databases. In those categories where software is both commercial and home-grown, such as finite element, computer graphics and CAD/CAM, one assumes that the commercial software is used in a different way to the locally written material. Tutorials feature heavily as in-house software. Interactive video and hypermedia have not yet caught on.
The X Window System Version 11 de facto standard has resulted in the appearance of a new peripheral known as an X terminal. Designed to support a graphical interface for networked applications, it is in the early stages of finding its home in the marketplace. It can be regarded as a development of the discless workstation, a system which runs technical applications using software loaded from a central system. However, X terminals handle only the user interface and networking, whilst the main part of the application is performed elsewhere on the network. In a sense, the X terminal is a replacement for the ASCII terminal, giving a graphical interface instead of the earlier character-based interaction. The program that drives the display in an X Window System is the X server. Programs that produce output to be displayed by the X server are clients. The X server and clients need not be running on the same computer. When X terminals are used to display the output, it is usually only the X server which runs on the X terminal. The clients run on another computer connected to the same network as the X terminal and communicating with it, usually via ethernet. The computer running the clients - the client server - could be a workstation or a separate compute-server on the network. However, the Dicoll terminal has both X server and X client running on the host, with only a minimal display system in the terminal itself.
Each X terminal has been tested with the clients running on an EASE Approved workstation (a SUN 3/60 with 8 Mbytes of memory) and on a more powerful compute-server (a SUN 4/280 with 32 Mbytes of memory). To provide a baseline, the tests were run on the same client-servers with the X server running on the SUN 3/60 - so for one set of tests, the X server and client-server were the same computer (this is how X programs would normally be run on a workstation).
This article presents a brief outline of the results. More detail can be found in the full report, which is available on request from RAL.
A total of 16 suppliers, a mixture of producers and distributors, were invited to take part in the exercise. Those that agreed to participate are listed in the table below, together with the machines supplied for testing (in most cases, the suppliers offer other models as well). Prices are based on list price, ex VAT, with the stated memory configuration. The key to the Price is as follows:
| Code | Price range £K |
|---|---|
| A | <1.5 |
| B | 1.5-2.0 |
| C | 2.0-2.5 |
| D | 2.5-3.0 |
| E | 3.0-3.5 |
| F | 3.5-4.0 |
| G | >4.0 |
One of the early producers of X terminals in the US was Network Computing Devices Inc (NCD) who quickly found several distributors, including a number in the UK. Some of these re-badge the terminal as an accessory to their other computing equipment. The Data General (DG) AVX-30 is a re-badged NCD16 and the Tektronix (TK) XN7 is a re-badged NCDI9. Princeton also refer to the Ultra 16 as the C-30.
| Supplier | Model | Memory Mb |
Monitor inches |
Price | Processor | Clock speed |
Max memory (Mb) |
Resolution hor*vert |
|---|---|---|---|---|---|---|---|---|
| DEC | VR150 | 1.0 | 15M | C | T138010 | 50.0 | 4.0 | 1024*864 |
| Data General | AVX-30 | 1.5 | 16M | C | MC68000 | 12.5 | 4.5 | 1024*1024 |
| Dicoll | Optima-X | 0.5 | 14M | A | MC68000 | 12.0 | 0.5 | 800*600 |
| Mellordata | X-15 | 1.0 | 15M | B | MC68000 | 16.6 | 4.0 | 1024*800 |
| Mellordata | X-19 | 1.0 | 19M | D | MC68000 | 16.6 | 4.0 | 1024*800 |
| NCR | XL-15 | 1.0 | 15M | C | MC68020 | 16.7 | 4.0 | 1024*800 |
| Princeton | Ultra 16 | 1.0 | 16C | C | T134010 | 50.0 | 8.0 | 1024*768 |
| Tektronix | XN7 | 2.0 | 19M | E | MC68020 | 15.0 | 8.0 | 1280*1024 |
| Tektronix | XN10 | 2.0 | 19C | G | Intel386SX | 8.0 | 1024*768 |
X terminals usually have two types of memory: ROM and RAM. The former contains the minimum program for startup. Some terminals keep most of their operating system in ROM, but the majority use it merely for the initial bootstrapping. Such terminals use much more RAM for the system. In addition, RAM is required to hold any fonts downloaded from the host.
With Dicoll, the X server runs in the host virtual memory so that the only memory required in the terminal is for font storage.
Most systems support X11 Release 3. DEC however supplies X11 Release 4. The fonts used to display text on the terminal come either resident in ROM or are downloaded from a host. Any number can be displayed simultaneously.
The method used to download fonts and the server, when not in ROM, is via tftp - the DARPA Trivial File Transfer Protocol. Unfortunately, standard tftp allows access to files on a remote machine without any authorisation checking, thus creating a potential security loophole. SUN and DEC provide a more secure version of tftp for this reason, which limits the directories which can be accessed by tftp. However some are not able to use the restricted version of tftp because the file locations demanded by them do not fit in with these restrictions. Others provide methods of altering the file paths so may be able to run under SUN's restricted version.
Most terminals use the ethernet for connection. Dicoll however offer only an RS232 serial connection. Most of the other terminals can also communicate over an RS232 connection, but this method is slower. The DG AVX-30, the TK XN7 and the Mellordata offer DECnet protocol as an option.
Benchmark programs have been obtained from various sources: previous assessments, public networks and specially written. Create Window and Graphics Context Switch represent some of the X primitive operations tested. The EASE tests are taken from those used in the workstation evaluations, and are fairly simplistic. This means that they generate much more ethernet traffic than the others, and this appears to be the reason why the results differ (Dicoll did very well because it was connected to a SUN 4/110). The Xstone test is based on the weighted combination of a series of measures, devised by C Gittinger of Siemens (this measure is often quoted in the Press). The remaining tests concentrate on those parts of the X system which allow the construction of pictures. Finally, a number of users were asked to try out the terminals, and give their subjective views on the experience.
The table presented below is an attempt to summarise the general characteristics and performance of each machine. This has been done by assessing each characteristic for each machine on a scale of * (poor) to ***** (excellent). A - indicates an apparent lack of the feature. The assessments are non-competitive, and do not represent a ranking.
| Supplier Model |
SUN 3/60 |
DEC | DG | Dicoll | Mellordata X-15 |
Mellordata X-19 |
NCR | Princeton Ulta 16 |
TK XN7 |
TK XN10 |
|---|---|---|---|---|---|---|---|---|---|---|
| Screen Size | *** | * | ** | * | * | *** | * | ** | *** | *** |
| Screen Size | *** | ** | *** | * | ** | ** | ** | ** | **** | ** |
| Screen Size | **** | ** | ** | * | ** | ** | ** | *** | *** | *** |
| Screen Size | - | ** | - | - | *** | *** | * | **** | *** | ***** |
| Screen Size | *** | ** | *** | * | *** | *** | ** | *** | *** | ** |
| Screen Size | *** | * | ** | ** | ** | ** | **** | ** | *** | * |
| Screen Size | *** | ** | * | *** | **** | **** | *** | *** | *** | **** |
| Screen Size | *** | ** | *** | **** | ** | **** | **** | ***** | **** | **** |
| Screen Size | *** | * | * | * | ** | ** | *** | ** | **** | * |
| Screen Size | **** | * | ** | * | ** | ** | *** | ** | *** | *** |
| Screen Size | *** | ** | * | * | * | * | ** | **** | *** | **** |
| Screen Size | ** | *** | ** | *** | **** | **** | ***** | ***** | ***** | **** |
| Screen Size | *** | ** | ** | ** | ** | ** | **** | ***** | **** | ** |
| Screen Size | *** | ** | ** | * | ** | ** | *** | ** | *** | *** |
X terminals are expected to be cheaper, quieter and require less administration. However, the hardware consists mainly of those components of a genuine workstation which are the most expensive, so the ability to keep prices significantly lower must of necessity be limited. As the assessment completed, at least one workstation supplier announced a discless, quiet workstation whose price is very competitive with all but the cheapest of these terminals. The functionality of such a workstation is obviously far greater than that of a terminal.
However, it is also possible to view an X terminal as an improvement on a standard ASCII terminal. It is usual to consider such terminals attached to powerful multiuser computers, which are capable of supplying the compute service needs of the more sophisticated device. This appears to be a much more cost-effective use of X terminals, providing multi-window capability to those who are used to a single screen only. This view seems to be borne out in practice, since several suppliers of high-end multiuser minicomputers also supply X terminals (often re-badged).
It is also possible to obtain software for Pes which will allow them to provide X terminal capabilities. This type of software is the subject of an evaluation currently being undertaken by the Computer Board.
The report was presented in June 1990. The committee decided to recommend that the NCR XLl9 (19in screen version of the NCR actually tested), Tektronix XN7 and Tektronix XN10 should be placed on the list of Approved terminals. The NCD 19 should also be placed on the list, being the same machine as the XN7.
However, CFTAG also recommended that those wishing to obtain X terminals should be advised that they appear more suited for use with multiuser machines, rather than in a client/server network. It is also necessary to take into account the potential problem with tftp.
There are two main strands of User Group within the UNIX community: commercial and academic. The commercial groups, such as Uniforum, consider UNIX in the marketplace, and concentrate on those aspects which concern the business world. The academic groups cover more detailed technical issues, often at the forefront of research. The European UNIX User Group (EUUG) is one of these, covering the whole of Europe (both East and West). Its definition is somewhat loose, since there were papers and attendees from the USA and Japan at their Spring Conference, held in Munich in April. The Conference proper lasted three days, with a couple of days of tutorials beforehand, and included a small exhibition. In addition, EUUG provides at cost a tape of public domain software and data images which cover a wide range of topics, both serious and trivial. This year's tape had a whole section on MacIntosh software as well as the usual UNIX collection. The graphical theme covered a variety of plotting programs, including a small ray-tracer, and image data from a variety of satellites (including Voyager).
I attended the Conference and exhibition, rather than the tutorials. The theme of the Conference was Object Oriented, which meant that there were a number of complex research papers detailing more Object Oriented language invention. As might be expected from an event held in Munich, over half the 400 delegates came from Germany. However, I was surprised at the shortage of UK representatives this time (approx 40); many of the major UK UNIX Universities did not send anyone (this contrasted with the last EUUG I attended, several years ago, in Florence). The French also were conspicuous by their absence (around 20 attendees).
One of the main issues at present in the UNIX world is the new Mach kernel from Carnegie-Mellon University, which is being incorporated into the Open Systems Foundation (OSF) version of UNIX (OSF/I). A paper was given on the main features of the system, which includes facilities to make the UNIX fork process much more efficient. In passing, the author commented that the name Mach had been chosen on the toss of a coin, and had originated from an Italian who had mis-heard an earlier suggested name!
It is not appropriate to go into detail on the papers here, since many of them would be of more interest to Computer Science researchers. However, there was a useful paper on NFS efficiency, which introduced the Prestoserv add-on board for SUN Servers. This is claimed to provide considerable speed-up in NFS writes, and hence improve the overall efficiency. Informatics have recently been testing such a board. There was also a surprising paper from a Frenchman about writing viruses in UNIX. Having seen that most of the virus publicity centres on PCs and Macintoshs, he set out (with his Management's approval) to see how easy it is to write a UNIX virus. Not only did he succeed, but he also let it loose on his unsuspecting colleagues!!! He then published the source code!! I hope that this does not encourage anyone in UK universities to try something similar.
As well as the planned papers, there were a number of so-called Birds of a Feather sessions, where groups with common interests met for discussion. The session on UNIX standardisation covered the current state of all the many standardising bodies relating to UNIX (now numbering 15!). There was a comment that, as things stand at present, one of the POSIX committees (1003.4) which considered Real-Time extensions, had recently taken on extra tasks not strictly covered by the title. Since however every item covered by the committee is at present considered optional, it is possible to conform to this particular standard by having none of it! Another session covered the current state of OSP. During this, it was pointed out that, while OSF are offering the source of OSF/I to Academics at a discount, it was at present necessary to have a AT&T licence as well in order to be able to run it legally. This problem is being addressed, and the link with AT&T code will be broken at some future date. There was also comment that the work on standardising binary interfaces (being undertaken - incompatibly - by both OSF and UNIX International) could well encourage the spread of UNIX viruses (such a standard being one of the reasons for the spread on PCs).
I found the exhibition a bit disconcerting, since it was aimed at the majority of attendees, and so was mostly in German! This included the available literature, and sometimes the internationalised Operating System interface as well. Among the exhibitors were IBM, who were showing their new RISC machines, and SUN, who had a stand but did not appear to have any English speakers manning it. A company called Servio Corp (USA) were demonstrating an object oriented database system called GemStone, which allows users to insert their own search functions. Non Standard Logics (French) were showing an X windows development environment called XFaceMaker, running on a Sony workstation linked to a SUN. The SUN had a French version of UNIX, so that the error screen talked about a rayon cosmique having hit the system! Of the two response buttons provided, one was approx 10 em long to accommodate the appropriate French phrase, while the other contained a phrase not capable of direct translation in a UK public Newsletter!
Anyone interested in obtaining more details of this conference can contact me for a copy of an internal report. More details of the contents of the EUUG tape can also be provided to anyone interested.
MMTACC came into being at an inaugural open meeting held in June 1989.
At this meeting the Club's objectives were defined and an executive elected charged with meeting these objectives. The Executive met for the first time on 25 September and allocated areas of responsibility and discussed ways of fulfilling the objectives.
During the past year the Executive has held three meetings and the first two of these took place at sites (Birkbeck and Keele) where parallel computing using Transputers was being developed.
At its last meeting the MMTACC Executive organised a session on Molecular Modelling at the Second International Conference on the Applications of Transputers (Southampton, 11-13 July 1990).
This has been set up as a subset of the main Software Exchange Library. Contributions are welcome in all aspects of molecular modelling, including energy calculations, simulation, structure prediction and visualisation.
A directory of people active in the molecular modelling area and using Transputers is being maintained at RAL.
A Bulletin Board has been set up, where abstracts of molecular modelling papers can be submitted.
On 19 September at the Rutherford Appleton Laboratory the MMTACC will be holding its annual Open Meeting, which will be followed by a seminar on Molecular Dynamics. Significant developments are taking place in the implementation of codes for molecular dynamics on parallel computers. The purpose of this seminar is to bring together some of the people in the forefront of this important area. Speakers will include:
Membership is free and is open to anyone interested in molecular modelling.
The North & North-West of Computing Short Courses - a collaboration between the Universities of Manchester and Glasgow.
Paisley College of Technology
MEDC offers courses to Scottish industry and lecturers from post secondary education.
IEE Vacation School at University of Strathclyde
2-7 September 1990
The School will show opportunities offered for expert systems and knowledge based techniques in control and will introduce the fundamentals of the subject. Software demonstrations will be presented with hands-on experience gained. Talks will include interface between control and expert systems, knowledge acquisition, software shells, rule based control techniques, industrial expert systems software and related control engineering topics.
7th September 1990, AIAI
Engineers often express concern about how they could integrate knowledge based systems with other software like databases, numerical analysis packages and CAD packages. This is because a large amount of effort has already gone into developing and maintaining these more conventional software. It would be undesirable and, in some cases, even impossible, to rewrite existing programs in AI languages or knowledge based system development tools.
The AI Applications Institute (AIAI), University of Edinburgh, organised a one day seminar on Integrating Knowledge Based and Conventional Systems in May. Due to popular demand it is to be repeated on 7th September 1990 at AIAI.
This one day seminar looks at the interface facilities provided by some of the current AI tools and languages. The systems covered will include Crystal, Goldworks, Nexpert, Leonardo, Edinburgh Prolog, and Xi Plus. The range of facilities provided by them varies between access to commonly used external file formats like Lotus 1-2-3 and dBase III to the ability to load in and execute routines written by the users in programming languages like Fortran and C. AutoCAD, a popular CAD package, with its own embedded version of Lisp - AutoLisp - will also be considered. Demonstration of some of these systems will be provided.
The provisional programme is:
This event is free to academic engineers.
The themes of ASE 91 will have a stronger emphasis on parallel algorithms for the efficient solution of partial differential equations, on examples of large scale computation which have had an impact on an engineering design, as well as on hardware and software aspects of supercomputing which result in more efficient indirect memory addressing, ie. for numerical solution methods on unstructured grids. Invited speakers on these and other relevant topics will be disclosed in a future announcement.
The review of Phase 1 of the Interactive Computing Facility (ICF) concluded that the concept of providing a distributed interactive facility for use by engineers as defined in the Technical Group Report was proving successful. Consequently a second phase covering the period 1979 to 1984 was proposed and approved. The overall objectives were:
To achieve these the Interactive Computing Facility had to:
The details of the plans for Phase 2 were widely distributed to the Academic Engineering Community and, in particular, Institutions were invited to submit bids for the provision of a multi-user machine, either a Prime or a GEC.
Successful applicants were provided with the appropriate machine and a six-year support contract. The contract covered the full cost of maintenance, the salary of a systems manager, travel and consumables for three years and then reduced linearly to cover maintenance only by the end of the fifth year.
Despite the additional case required to justify the provision of a Prime machine, five successful bids were made in 1979 and the first remote Prime, a P550/1, was installed at Sussex University in October 1979. Four further P550/1 machines were installed at City University, Surrey University, Warwick University and University College, London in 1980. Additionally an existing P300 machine at the University of East Anglia was upgraded to a P400 in 1979 and a second Prime service machine, a P400, was installed at RAL in June 1979 to serve as a development machine and provide a service to local users.
Twelve GEC 4000 Series machines were installed in the period 1979 to 1981 at the following sites:
There was also a review of the DEC10 KI services at Edinburgh and UMIST in early 1980 which resulted in the machine at Edinburgh being upgraded to a DEC 10 KL and the UMIST machine being replaced by a P750. UMIST also took on a responsibility, under contract to SERC, to provide user support for the Prime community.
The remainder of the Phase 2 period saw a steady consolidation of the Prime hardware and the associated systems and applications software. A planned programme of memory, disc and processor upgrades reflected both the increasing number of users of the Prime Service (1000 by 1981) and their increasing demands on the computing resources. Further development of PRIMOS concentrated on the provision of X25 networking to ensure that all the Prime machines were connected to the SERC X25 Network.
The size of the terminal pool had been increased to 1000 (50% being graphics terminals) by 1982 and these were widely distributed to grant holders throughout the UK. In many cases terminals were connected to remote machines via time-division multiplexed leased lines.
By the end of 1981 there were ten Prime systems in the ICF, 3 × P750, 5 × P550 and 2 × P4OO. The expected number of simultaneous users which could be supported were:
The good experiences of the ICF with Prime machines resulted in a number of other projects purchasing Primes (details of these are given later in this series).
All these systems were connected to the SERC X25 Network and offered both interactive terminal and file transfer facilities. Support and distribution of the PRIMOS operating system was carried out via the network. All Primes were able to use the file transfer facility to access the large central IBM facility at RAL for computationally intensive activities.
The AP-120B was transferred to the P750 at RAL when it replaced the original P4OO. Although a few users made effective use of this array processor the overall assessment concluded that there was great difficulty in using the machine because of inadequate software and the AP-120B was taken out of service in 1983.
The applications activities on Prime during Phase 2 concentrated on Control, Electromagnetics, Finite Elements, Digital and Analogue Circuit Design, Pre- and Post- Processors, Graphics and Database Management Systems. It had become clear by 1981 that the Primes were far more effective than the GEC machines for running real applications (thereby confirming the recommendations from the assessment programme in 1977). GEC users requiring to use the Control, Electromagnetics, Finite Elements and Circuit Design packages had to access a Prime over the Network.
The SERC version of PRIMOS was stabilised during this Phase of the ICE A major problem of making such major changes to the standard PRIMOS was that it took at least 6 months to incorporate them when a new Revision was received. In practice this meant that the ICF did not take every Point Release of PRIMOS. Several unsuccessful attempts were made to get Prime to accept the SERC modifications during this period and thereby provide support for the ICF operating system on the Prime machines. Needless to say there were a large number of demonstration disputes with Prime over the years!!
New software additions to all the Primes in 1984 included PROLOG, EMACS and the 'C' Compiler. A new element appeared in the computing facilities offered by the Engineering Board to its users in 1982, namely the planned introduction of Single User Systems (ie workstations) as part of a Common Base Policy. By the end of 1984, these systems were beginning to reduce the demand for ICF resources. Strategically, from SERC's point of view, Single User Systems were now the 'state of the art' computing systems and Multi-User Minis were now 'run of the mill'.
A further problem in 1984 was that there were severe financial pressures on SERC which were expected to remain for several years.
The last article in this series will cover Phase 3 of the ICF (1985 to 1988).
