In response to the requirement identified at the CFTAG Workshop the Computing Facilities Committee and the Computer Board have recognised the substantial needs for provision in computer graphics, which have only been partially met by Uniras products made available via CHEST. The Board, together with the Computer Facilities Committee of the Engineering Board of the SERC, has therefore decided to fund a coordination post at Loughborough University together with a small group, Advisory Group on Computer Graphics (AGOCG), to help the funding bodies develop policy by providing appropriate advice.
To advise the Computer Board and Research Council on all aspects of computer graphics.
To be aware of the advances in computer graphics in both the standards area and in innovative new technology in both hardware and software.
To liaise with the community to help identify requirements in the area of computer graphics.
To recommend to all the relevant funding bodies options for purchase, support and development that would improve the environment available to the academic community in the area of computer graphics.
To ensure that facilities for education and training are provided to the community on the benefits arising from the use of computer graphics including standards.
The Group will initially formally report to the Board and CFC annually. In addition, matters for consideration will be referred to it by the Board and other bodies from time-to-time.
Obvious ways in which the Coordinator and Group will fill their role of helping identify community requirements will be as a result of feedback from existing groups such as IGWP and information and advice on the experiences of users. If you have input, please contact Dr Anne Mumford at Loughborough.
The Computing Facilities Committee has approved the sponsorship of a small number of places at each of the EASE Education events for the next year. The sponsorship covers travel and subsistence only, except in the case of courses, etc run on behalf of EASE by the Artificial Intelligence Applications Institute, Edinburgh where the appropriate course fees are also included. Details of the Education Programme are published in the Engineering Computing Newsletter.
To apply for one of the sponsored places, a short case should be sent to: Mr G A Lambert, Informatics Department at least six weeks before the actual event you wish to attend. The case must be signed by your Supervisor or Head of Department. You will be notified of the decision on your application at least three weeks before the event takes place.
Throughout the 70s considerable advances were made in the application of computers to the prediction of the cost and performance attributes of buildings at the design stage. For example, at the University of Strathclyde the ABACUS and ESRU research units have, with SERC funding, constructed several modelling systems: for the simulation of the flow of energy (ESP) and light (DIM); and for the real-time visualisation of a building design within its environmental context (CAVIA). By the early 80s these, and other systems, were beginning to emerge as viable, if not user-friendly alternatives to the range of simplified design tools then in widespread use.
These emerging simulation systems were characterised by
At the time only mainframes were perceived to be capable of satisfying these requirements and so several machine implementations had to be supported with all the concomitant resource implications. By 1984 the situation had changed dramatically with workstation technology reaching a stage in its evolution where favourable cost/performance ratios were being delivered at an acceptable cost. The conversion of a model such as ESP was identified as an important step in the technology transfer process. This article chronicles this conversion and the further developments then enabled.
The first task was to install ESP on a Unix workstation: a SUN2 with 4MBytes RAM and 170 MBytes hard disk was the target machine. Because of the SUN's non-sharing libraries this resulted in a 25% increase in memory requirements and, because of the modest power, annual simulation times rose markedly to around 45 CPU-hrs. Bit-mapped graphics presented an initially insurmountable problem which necessitated a decision to utilise Tektronix emulation in the short term. In the medium term a move to OKS was contemplated but was rejected on the ground that since OKS is not raster-based, it could not be used to produce good interfaces. In the event the WW graphics toolkit from RAL was selected. WW is an open-architecture, raster-based system designed to work with X-windows. As a lowest common denominator, Unix based system, it offered the best portability: it already operated on three Unix workstations for example and could be fairly easily ported to others. The effort to convert ESP to utilise WW, of the order of 6 person-months, has paid dividend: currently ESP is being endowed with an intelligent front end which operates under X-windows.
Another time consuming task involved the retro-fitting of the ESP source code to enable the dynamic assignment of all I/O channels as a function of the run-time context. By allowing external processes to control the source and destination of all user I/O, file I/O, graphics and error handling, jointly and severally, new opportunities were created for automating entire performance assessment methodologies by the introduction of generalised Shell Scripts, each parameterised in terms of the context dependent knowledge required for their execution.
The benefits to accrue from this conversion of ESP to run under Unix were essentially threefold. Firstly, the strong file typing of Unix brought a more coherent structure to ESP and helped to transform the system into a powerful machine environment. Secondly, ESP's graphics capabilities were markedly improved. Lastly, the parameterised methodology scripts enabled users with no prior knowledge of ESP's operational syntax to gain access to the power of a simulation based approach to performance prediction at the design stage.
By 1985 ESP was firmly established under Unix and the research focus shifted to address new issues. Two development areas in particular were fostered by ESP's new machine environment: the introduction of an Intelligent Front End (IFE) and the development of an object orientated (00) approach to building environmental simulation.
The objective of this project was to develop a machine environment which would allow designers to describe a building and initiate performance appraisals in a manner which depends on their design perspective and level of computer expertise. The IFE comprises a central communications module and several cooperating programs (running asynchronously) which can examine this module for information and post results to it. Each program is mono-functional: addressing dialogue, knowledge, user, data, appraisal and application program handling. The lFE is therefore an intricate synthesis of user handling, human-computer interface techniques, contextual knowledge and the application interface at its back-end. The system is designed to manipulate different user conceptualisations. where each conceptualisation is conceived as a collection of related concepts matched to a set of Prolog knowledge bases for concept value validation, inferencing and user dialogue. These conceptualisations are, in turn, manipulated by a forms package running under the control of a further set of Prolog predicates.
The IFE project, which is funded by SERC and involves a collaboration between the University of Strathclyde and the Rutherford Appleton Laboratory, is due for completion by July 1989. The next stage will be to implement a number of field trials to explore system performance in practice.
The EKS project has only recently commenced as a SERC-funded collaboration between the Universities of Strathclyde and Newcastle, Cranfield Institute of Technology and Rutherford Appleton Laboratory. It is an attempt to utilise the 00 programming paradigm in the development of an advanced machine environment for the construction and maintenance of building environmental simulation models in general. The aim is to identify the spectrum of computational methods which underpin such models, to develop an appropriate object class structure for the containment of these methods and to install these objects within an 00 database system. It is anticipated that such an environment will encourage better code sharing among researchers, facilitate more rapid prototyping and help to establish validation centrally within the model building process.
In tackling these projects several problems have been encountered and some insight gained into the issues which will confront the further evolution of a simulation based approach to CABD. These are:
In moving from the initial SUN2 machine to a SUN3/60 the speed improvement for a system such as ESP is significant: around a factor of 3. This means that a typical thermal simulation of a year in a building requires in the order of 15 CPU -hrs. This figure is further reduced by a move to RISC-based machines (around 2 CPU-hrs on a Mistral Hi-Techl0 based on the MIPS R2000 VLSI RISC chipset for example) bringing closer the possibility of building/plant control systems based on real-time simulation. Nevertheless, the IFE system, where several CPU-intensive programs must co-exist, has demonstrated the need for substantially more computational power. This could be achieved through parallelism and distributed networks of high MIPS nodes or by an application-oriented transputer environment.
The EKS project is an attempt to utilise an 00 approach to model construction. To this end 00 Languages (CH) and 00 Databases (Vbase) are being used. It has been found that these systems are not particularly appropriate for use on engineering design applications.
There is a distinct need for better toolkits to overcome the present situation where application authors are forced to produce customised interfaces utilising low-level facilities. Such toolkits are appearing but an effort is required to identify the best of these in the context of engineering design and to promote their use through, for example, inclusion under EASE.
Within the IFE project several AI problems are being tackled which would suggest possibilities for the future. In particular there is a need to undertake more research in the field of user conceptualisation and user modelling, to develop more efficient knowledge representation languages suitable for the manipulation of engineering design knowledge and to make data storage and retrieval more autonomous and intelligent.
In conclusion, the advent of the workstation has made a significant contribution to the transfer of CABD into practice. At the same time it has opened up new possibilities which clearly demonstrate that the evolutionary vista is a long one.
People who use Unix computers will quite likely already be aware of the Usenet News system, and the benefits it provides. The News system is a worldwide electronic bulletin board, run between co-operating Unix sites. It provides a forum for discussion of just about every conceivable topic. It is organised into 'newsgroups', each newsgroup dealing with a specific interest area. Users of the system subscribe to the news groups that are of interest to them.
The more popular topics include discussions on current computing technologies with news and views from a global user community; notices of forthcoming events both in the UK and worldwide; advertisements for and distribution of a large body of public-domain software; and a host of subsidiary topics of philosophical, sociological and cultural interest. In addition the system is an excellent forum to get hints and solutions for a computing problem -the community of subscribers is estimated at 400,000 worldwide.
The EASE programme has recently asked the UK News administrators to set up a news group within this bulletin board to disseminate information about EASE activities and associated issues. The news group is called uk.eng.apps. Besides providing a forum for announcements, it will allow people in the EASE community the opportunity to discuss ideas and topics that might be of interest to the wider community.
If your site does not already receive this News and you have a networked Unix system, then we have the ability to forward (or feed) the News to a limited number of 7 sites. If you are interested in doing this then please contact Neil Calton.
One word of warning though - such is the quantity of information available that, unless you are prepared to be very selective in the number of newsgroups you receive, you will need to allocate around 30 Mbytes of disk space on your system. Receiving the News is a task ideally suited to one of the EASE Central Servers now being installed in Departments around the country.
We love CGM! This was the message on the stickers being given out at the Eurographics-UK demonstration of the CGM at Manchester University 29-31 March 1989. This event gave the opportunity for the SERC to make their community more aware of the importance of the CGM standard following the Computing Facilities Technical Advisory Group (CFT AG) workshop on computer graphics standardisation which noted its potential in the community.
The CGM is an international standard which specifies a file format for the storage of pictures on computer systems. It is therefore a non-proprietary format. Pictures can be stored in CGM format in a business graphics package and transferred to, say, a graphic arts package, a slide making service or a desk top publishing system.
That is the theory and at Manchester University we saw this put into practice. Over 20 companies cooperated to get this CGM demonstration off the ground. The event was coordinated by Dr Anne Mumford of the Computer Centre at Loughborough University. This was the first event of its kind in Europe and the first in the world to show the full use of the different encoding techniques of the CGM.
The demonstration revolved around two fictitious companies who reflect the needs of the many organisations.
The first saw a company called Eurocorp which has been set up with different departments in different countries in anticipation of the single European market in 1992. The company is in the aerospace industry. The task before us at the demonstration was to produce a report detailing a new advance in technology; to produce some slides of the information for a presentation and to produce some publicity material. The demonstration showed the initial inputs of a logo. coming from Eidographics/GSC, an engineering drawing from MicroCADAM and some flight path information from BUSS Ltd. The flight path information was extended by the GTS-GRAL software and passed on to the Ventura Desk Top Publishing package together with the engineering drawing. The pictures were slotted into a report which was printed out on the Mitsubishi G650 colour thermal plotter. Our slide was produced via the Pansophic Studioworks package and the poster was produced by the Computer Associates Super Image package and output on a Precision Image electrostatic plotter.
The second tour showed a consortium of companies bidding for the channel tunnel project and having to show the Government representatives that they could react quickly and in an integrated way to requests for information. We saw a visualisation of the Channel Tunnel terminus using the SAMMIE CAD robotics package. The results were a report produced on a Mitsubishi plotter from Ventura; slide output from Letraset; some promotional material produced using Ginoslide; and some proofs of the graphical data for discussion purposes including some geological sections produced by Uniras.
The Central Computing Division of the Rutherford Appleton Laboratory have developed a system for putting CGMs together in a video. Participants in the demonstration sent metafiles to the Laboratory and the resulting video was shown.
It is essential that metafiles conform to the standard and testing facilities were on show at the exhibition. The National Computing Centre and System Simulation have worked together to develop a CGM checking facility. Another package from the US-based company CGM Technology Software was also shown.
One of the messages coming out of the demonstration was that suppliers should not be afraid of using the CGM standard. By including it as an option for importing and exporting graphical information software will be enhanced. Purchasers of graphics software should be demanding CGM when considering software packages.
The Computers in Teaching Initiative (CTI) was launched in 1984 by the Computer Board for Universities and Research Councils and, in conjunction with the University Grants Committee, funded 139 projects to promote and develop the use of computers in undergraduate teaching. One of the 15 engineering projects, the Computer Aided Design in Education (CADEd) project, ran at Queen Mary College until 1988. Through CADEd, approximately 35 existing packages used in teaching engineering were ported to 32-bit workstations, 11 new packages were developed and documented and several were entered into the Engineering Science Program Exchange. In January 1989, 20 CTI Centres were established and the CTI Centre for Engineering is now under way at Queen Mary College, University of London to provide a service to engineering departments around the country. Its principal objectives are:
We shall be communicating with all engineering departments in the UK in the coming weeks, and details of this activity will be the subject of regular articles in this Newsletter.
SERC/Alvey infrastructure funding for POPLOG has now ended, so it will no longer be possible to provide POPLOG free to SERC grant holders. However, the University of Sussex will attempt to keep prices for UK academics as low as possible.
Development work is continuing on all the main POPLOG languages (Prolog, Pop-11, Common Lisp and ML), the X11 window interface and new tools. Additional ports are under discussion with hardware manufacturers. We shall attempt to continue to support a very wide range of POPLOG users in teaching, research and development, in fields as diverse as ship design, natural language interaction, Image interpretation, text processing, real time control and support for novice programmers.
For POPLOG commercial users the marketing arrangements have been changed. SD-Scicon, with the agreement of the University of Sussex, has transferred POPLOG rights to a new company, consisting of staff formerly responsible for support and distribution of POPLOG to commercial customers, and we are pleased to announce that they will be responsible for all commercial and overseas sales and support of POP LOG. Enquiries to: Colin Shearer, Integral Solutions Ltd.
The company will also be announcing new products, some based on POPLOG.
The new breed of terminals has emerged within the last year which implement the MIT X Window System software and display graphics-rich software on high resolution screens. These X terminals are seen as a low cost entry point for Unix users wishing to access software located on various computer systems around a network. They have certain workstation-like features, such as a mouse and bit-mapped display, and can access much of the software presently used on a real workstation running X, but at a much lower cost. They also tend to be cooler, and quieter, than a workstation.
The X Window System provides a base for Window systems on Workstations, and is becoming generally available across a wide range of different manufacturer's machines. It separates windowing into two parts: the application (or client) and the display station (or server). These two parts do not necessarily have to run on the same machine. With an X terminal, the server runs in the terminal (either being located in ROM or being downloaded from a host) and the appropriate client runs on the host. The X server controls access to the display by various applications, interprets network messages, passes user input to the various applications, performs two-dimensional drawing and maintains complex data structures such as windows, cursors and fonts. Support for backing store and off-screen graphics reduce network traffic by storing images until they are needed again.
There are several devices on the market at present, and more are due to appear this year. As part of the EASE activity and in collaboration with the IUSC Working Party, it is intended to find out as much about these terminals as possible, and to compare the different characteristics. Results will be made available via the Newsletter.
The increasingly widespread use of Sun workstations has generated a need for training SERC grant users in the driving of a Sun workstation, giving them an overview of SunOS (UNIX), and providing an introduction to the software available for developing applications under SunOS. A New User course is being developed at UMIST to deal with the following topics:
The first course is likely to be given at the Rutherford Appleton Laboratory during July/August, but subsequent courses could be held on a regional basis if there is a demand.
It would be helpful to the planners to have some feedback from prospective attendees about their requirements, particularly with regard to the demand for basic UNIX instruction.
The Artificial Intelligence Applications Institute (AIAI), University of Edinburgh, is a technology transfer unit, facilitating the use of AI tools and techniques on real problems by communities that are not AI specialists.
As a part of the EASE programme, AIAI is working on a project, funded by the Computing Facilities Sub-committee of the Engineering Board of the Science and Engineering Research Council, to provide support to SERC funded engineers who wish to investigate the application of AI in their work.
Research staff and students who already hold or who are working on grants provided through the SERC Engineering Board may visit AIAI to use specialist hardware and software systems. Approved pump-priming access to these systems is also available.
The Visitor Scheme offers various services:
AIAI hosts a unique UK facility known as the Knowledge Representation Systems Trials Laboratory (KRSTL). The KRSTL facility, which consists of general purpose and specialist AI computer hardware and software systems, is available for use by the visitors.
Experienced AIAI staff give technical support to visitors by giving advice, supervision or formal training sessions.
Our visitors find this scheme most valuable in helping them to make progress in the early stages of their projects.
Our short courses for the fourth quarter of 1989 are listed in the following table:
These courses are very popular, so early booking is advised. Provisional bookings on the courses should be made as soon as possible. As our support contract expires at the end of September and a proposal for continuation is currently being re-negotiated, their final availability will be subject to the terms of whatever contract is in place come October.
At UMIST, we are gathering together information and software which will facilitate the transfer of software from a SERC Prime (past or present) to a Sun. The team at UMIST, which formerly supported SERC Primes, intends to:
The UMIST team would appreciate contributions from users who have successfully transferred programs from a Prime to a Sun. The contribution may be in the form of information or software tools that they have used or developed. Similarly, users needing advice are asked to contact the UMIST team via E-mail at support@umist.
To date, based mainly upon the work of Russ Lomax and Alan Hulme at University of Warwick, the following tools have been collected:
IUSC WORKSHOP - 4-5 July 1989 The next IUSC Workshop will be on Algebraic Computing. It is being hosted by the University of Liverpool on 4/5 July 1989. The workshop will cover many aspects of Computer Algebra Software including:
It is hoped that demonstrations of systems such as Reduce, Maple, Macsyma, muMATH, Derive and Mathematica will be available.
London & SE Regional Transputer Support Centre Programme for May to November 1989
Transputer Awareness 25 July, 26 September, 2 November
Exploiting the Transputer 18 July, 13 September, 16 November
APPLICATION OF TRANSPUTERS CONFERENCE AND EXHIBITION, University of Liverpool, 23-25 August 1989. The first International Conference on Applications of Transputers will be held at Liverpool University, 23-25 August 1989. The conference will highlight advances in applications software, both in numerical applications (CFD, FE, Molecular Modelling, etc) and non-numerical applications (image processing, robotics, etc). Accompanying the Conference will be the largest Exhibition so far of Transputer products with suppliers from the UK, Europe and the USA being present. The Conference and Exhibition organised by the SERC/DTI Transputer Initiative is to be preceded by a half-day tutorial on Transputer Applications aimed at senior managers and researchers and designed to being out the likely future impact of Transputers.
EUROGRAPHICS CONFERENCE, EG89, Hamburg, 4-8 September 1989.
The theme for this year's Conference IS Interaction, Integration and Visualisation. The Conference starts with 2 days of tutorials on topics such as Visualisation of Scientific Data, Advance Topics in Solid Modelling, Distributed Window Systems, Object Oriented Graphics and Produce Data Exchange. There are parallel streams through the last 3 days of the Conference plus an Exhibition. Invited speakers include Jim Foley, Gerald Murch and Don Greenberg from the USA.
COMPUTER INTEGRATED PROCESS ENGINEERING CONFERENCE (CIPE '89), Leeds, 25-28 September 1989
CIPE '89 will be the foremost event for process engineers involved in almost any aspect of software development and use ever to be held in the UK. Over 50 papers from Europe, USA and Japan are currently being refereed.
UK IT 1990 CONFERENCE, University of Southampton, 19-22 March 1990
The aim of the Conference is to provide an annual, national, technical forum for the presentation of current work in the enabling techniques for information processing; ie to cover the middle ground between specific application areas on the one hand and basic research on the other.
1989 AIAI SUMMER SCHOOL ON THE APPLICATIONS OF ARTIFICIAL INTELLIGENCE, Model Based Reasoning in Engineering to be presented by John Kunz Chief Knowledge Systems Engineer, IntelliCorp and Consulting Professor of Civil Engineering, Stanford University at the Artificial Intelligence Applications Institute University of Edinburgh, 31 July - 4 August 1989
The Summer School will investigate issues concerning model based reasoning systems in engineering. In engineering and science, models are descriptive (of the concepts of a system, their attributes and their relations), and they are predictive (of the values of dependent system attributes following change in independent attribute values). Model-based reasoning systems use Artificial Intelligence techniques in the creation, simulation and testing of engineering models of complex systems. We will discuss such systems as extensions of Artificial Intelligence expert systems, consider methodologies for creating such systems, and evaluate their uses and limits in engineering.
