Douglas Lewin, Professor of Computer Science and Information Engineering at the University of Sheffield, died on 20 September 1988 aged 57, following a short illness.
He first obtained employment as an electronic engineer with Unilever and Marconi, before turning to an academic career at Brunel University. Here he carried out research into logic systems design and the use of computer tools for the digital design process. A lectureship in Digital Systems at the University of Southampton was followed by a Chair in the Department of Electrical Engineering and Electronics at Brunel University. The Anglia TV Chair of Electronics at the University of East Anglia was his next appointment, before moving to Sheffield in 1986.
He was prominent in the affairs of the Institution of Electrical Engineers, the Institution of Electronic and Radio Engineers and the British Computer Society. He also made his mark on a series of Science and Engineering Research Council computing committees, all of which in time he was to chair. His presence had a major effect on national policy in this field and he was looked on as a sage figure, still possessed of that youthful zeal which had always been his hallmark:. He was chairman of the SERC's Computing Science and Computing Facilities Committees (CFC) and a member of its Information Engineering and Computing Advisory Committees.
As chairman of CFC, Doug was responsible for overseeing the Engineering Board's computing infrastructure and, in particular, for the early decisions relating to the new Engineering Applications Support Environment (EASE) policy. He also chaired the Working Party in 1986 which recommended the establishment of the joint SERC/DTI Engineering Applications of Transputers Initiative. More recently, in 1987, he was instrumental in producing the successful joint bid by the University of Sheffield and Sheffield City Polytechnic to be the National Support Centre for this Initiative.
Doug's forward-looking perceptive and always realistic advice permeated many areas of higher education and industry, and his wisdom will be sorely missed by his many friends and colleagues.
The Computing Facilities Technical Advisory Group (CFTAG) has asked that the newly emerging computers based on the 80386 chip should be assessed. In order to perform this task, it is necessary to discover where the Community see this type of machine fitting. There are a considerable number now being sold, and many of these are able to run both MSDOS (like a PC) and UNIX. In performance, machines can overlap the low-end systems in the Workstation range (such as the Sun3).
I would be very grateful if you could help in formulating a specification. In particular, please let me know if:
If you are considering an 80386 machine as a large PC, I would like to know what PC applications software you want to run.
At present, the standard text processing software supplied with EASE-supported workstations is based on AT & T's Documenter's Workbench (troff etc). There are, of course, other systems available. One of these is TEX, with the associated LATEX. As well as the consideration of WYSIWYG document systems (see July-Sept 1988 issue of the Newsletter), there is a desire to see whether there are enough people in the Community requiring TEX to add this to the list of supported software. I would be grateful if those who would like to see this happen contact me (preferably by email -ret@uk.ac.rl.ib) so that the size of the Community can be estimated.
For the past four years Rutherford Appleton Laboratory has been a partner in ESPRIT CAD Interfaces (CAD*I), a project which was described in this Newsletter, issue 3, May/June 1987. RAL staff have been involved in the development of neutral file specifications for exchanging product analysis data between CAD systems and finite element modelling programs, between modelling and finite element analysis programs, and for converting analysis results into neutral file form for post processing. A body of software has been developed for writing and reading the files.
RAL has sent representatives to meetings of the International Standards Organisation Committee TC 184 SC4; this Committee is developing a product data exchange standard known as STEP (STandard for Exchange of Product data). Since CAD*I work on product analysis data is in advance of work being done elsewhere, it has been possible to have a strong influence on STEP. The first version of the standard is expected to be published this year.
Although RAL's work in CAD*I will end on 31 October this year, there is still much we can contribute to the development of STEP. Following the recommendations made by the recent Computing Facilities Technical Advisory Group (CFTAG) Workshop on Integration, we shall continue our participation in the STEP meetings.
The CAD*I involvement leads naturally to other tasks which have been recommended by the CFTAG Workshops. We are beginning a programme of assessment of database systems and their usefulness to engineers for storing information such as product model data. Currently, there is much interest in systems which will integrate the various stages of the design process, such as CAD modelling and finite element analysis. The work on analysis data exchange can provide a basis for developments which will keep the UK at the forefront in this field.
We have begun to formulate a programme which will extend the data exchange techniques developed for CAD*I and which, together with the database work, will be directed towards integration of the design process. Experience in CAD*I has convinced us that we need collaborators from other organisations in order to achieve success. If your department is interested in product information exchange, please let us know so that we can discuss our mutual interests.
The possibility of holding a workshop on data exchange is being considered. This would provide an opportunity for us to tell the community what we have achieved in CAD*I, and for the community to reach a consensus about what needs to be done.
An introductory article on ECSTASY appeared in the March/April 1987 issue of the Newsletter. This updates that article and describes the current status of the project.
ECSTASY has been under development over an 18 period by a team in the Control Systems Centre at UMIST and the prototype version will shortly be delivered to the Rutherford Appleton Laboratory for support and further development.
ECSTASY is structured to provide the core facilities for an integrated control system design environment. It is intended to meet three requirements:
The essential components of ECSTASY are shown below:
The mathematical basis of control system design is matrix manipulation and ECSTASY uses a proprietary version of MATLAB called PRO-MATLAB to perform this function. Also included are the specialised packages, Control Toolbox and Identification Toolbox, providing a useful range of control system design algorithms. Selected optimisation and other routines from the SLICE library are also interfaced to PRO-MATLAB.
Proprietary non-linear simulation packages are being incorporated within ECSTASY. ACSL, used world-wide for plant and process modelling, has already been interfaced by UMIST. TSIM, more recently developed by the Ministry of Defence for aerospace system modelling, is currently being interfaced by RAL.
ECSTASY has been developed using the Software Engineering Toolkit from PA Management Consultants.
The PA inSET module provides a development environment designed to allow the rapid prototyping and testing of the user interface and database structures. inSET provides a variety of input dialogue mechanisms. These include button, pop-up menu and forms. Window handling is provided by the module windowSET which currently allows 22 windows, which can be panned and zoomed independently, within the ECSTASY environment. The windowing system provides the following tools which may all be present on the Sun workstation screen at the same time:
The usual Sun window-based tools, such as text editors, shell tools, command tools and the Framemaker report generation software can be available at the same time as ECSTASY is active, since the ECSTASY environment uses only one Sun window.
Graphics facilities include 2-D lin-lin, log-lin, log-log and polar plotting and also 3-D mesh surface graphics. The 2-D facility includes control system specific plots such as Bode graphs, Nyquist diagrams, root locus plots and time responses.
The database structure has been designed by UMIST to meet the special needs of control engineering where the data objects are complex and have complex interrelationships. All data entities within the database have associated with them date stamps, access protection and relational pointers. A History function is provided enabling chosen series of past commands to be repeated at will.
A Linear System Manipulation facility has been incorporated, using a terse algebra to describe control system structures. The resulting linear models can then be transformed between a complete range of forms including transfer functions, state space description, frequency and inverse frequency responses and pole zero description and their sampled data equivalents.
A Non-linear System Manipulation facility is provided whereby a simulation file is created. This can be translated, compiled and run by the non-simulator programs, ACSL and TSIM.
Ease of porting is expected to ensue from the use of the onSET module of the SET Tools. This buffers the implementation from the machine operating system.
The software developed by UMIST is written in C.
ECSTASY currently runs under UNIX on Sun3 systems equipped with floating point co-processors. About 10 Mbytes of disk space is required by the environment. A trial port to Microvax II under VMS is currently in progress.
During the next six months ECSTASY will be undergoing user trials by a number of groups, each comprising academic and industrial staff. A range of engineering applications have been chosen, each having a strong control system design requirement Initially, these trials will be based on Sun3 (UNIX) but it is expected that later trials will use VAX/VMS systems.
To meet the objective of creating an integrated, rather than a federated design structure, it will be necessary to incorporate the novel design algorithms arising from academic research, rather than merely interfacing complete packages which may have been built around the novel design algorithms. In this way the common man-machine and database facilities of ECSTASY should be used to greatest effect, giving a common user interface to all design tools. Selection of appropriate algorithms for incorporation and the setting up of mechanisms to resource the integration work are issues yet to be decided upon by the SERC Computing and Design Techniques for Control Engineering (CDTCE) Management Committee, which oversees the project.
If industrial users are to be encouraged to use ECSTASY it must become available with accepted commercial support and maintenance. To this end discussions are being held with specialist software houses who may be interested in turning ECSTASY into a commercial product.
Ref. Computing and Design Techniques for Control Engineering - A UK Research Initiative by N Munro and P Hammond Proceedings of [MACS 88. 12th World Conference on Scientific Computation. Paris. July 1988.
The Computing Facilities Committee (CFC) of the Engineering Board of SERC are currently funding a three year contract at the AI Applications Institute, Edinburgh University, to provide support to engineers wishing to use AI/IKBS. This article describes some of our current activities and services available.
There has been a good response to our announcement in the last issue that a bibliography of AI in engineering and a list of AI tools are available for copying over JANET from uk.ac.ed.aiai. The tool list is continually being updated and has now been split into several separate meso The me index contains a list of the files available for copying by ftp. For example, assuming you use hhcp for ftp, try the following:
Home% hhcp uk.ac.ed.aiai:' <AILIB>index' index transfer authorisation: guest transfer password: <enter your user name here>
Those of you who use the VMS utility transfer, should remember to specify the switch /format=line to ensure successful copying. We've tried to make the file transfer process more robust, file names may be in either upper-case or lower-case, and the files are now purely textual, they don't need to be processed by any text formatters. However, we would remind you to use your user name as the password, that way if there are any problems we can get in touch with you via electronic mail.
Our next workshop, which will be on Artificial Intelligence in Process Engineering, is scheduled for 9 December 1988. There will be talks covering a variety of systems which involve the use of AI:
The systems described in the talks will be demonstrated and there will be ample opportunities for discussion.
Following the format of our previous workshops, we would like to limit the number of attendees to around thirty. If you are interested in attending, please let us know as soon as possible. The workshop has been scheduled so that those interested may attend the Expert Systems Course on the two preceding days, 7 and 8 December (see below).
In Issue 7 of the Engineering Computing Newsletter we described the Fundamental and Advanced AI courses that are available free of charge to researchers and students who are working on SERC funded projects. The following table gives details of courses to be given at AIAI during the remainder of 1988 and for the first quarter of 1989:
Please let us know of your interest as early as possible as some courses are fully booked long in advance.
A binary distribution of MIT's X Version 11 Release 2, including Xlib, Xt and other toolkits will be ready in October 1988 for Sun3 users wishing to use Xlib directly or use Xt (X toolkit). It includes the available fixes from MIT and some speed ups for the Sun X server. It will be available on the Janet Access Machine, when that becomes operational. If you have no JANET access you can get the software from ECF support. At present the source of Xlib, Xt, Andrew and InterViews is available from Imperial College.
ww will be available on top of Xlib from RAL. An X workstation driver for RAL GKS will be available. An object oriented toolkit may also be developed.
A Sun user note is being issued with fuller details of available X software. More information is available via the Service Line at RAL. Further news of X and related topics will be reported in this newsletter.
The Engineering Computing Facilities Executive (ECFE) is tackling the problem of making a list of applications software which should be incorporated into the Schedules A (Packages) and B (Libraries) of the Engineering Applications Support Environment (EASE) definition. Contact is being made with the Subject Committees of the Engineering Board to identify those items of software which should be included.
Suggestions as to what should be included are also welcome from SERC-supported grant holders, and these should be passed on to any member of the ECFE at RAL as soon as possible, as it is hoped to produce an initial draft proposal early in 1989.
Workshop on Al Support for Engineers, 13 December 1988 Rutherford Appleton Laboratory Lecture Theatre, Building R22
Engineering Computing User Group Meeting, 19 January 1989 Rutherford Appleton Laboratory Lecture Theatre, Building R22
Presentations include items on workstation and central server assessment, the EASE future programme and details on porting applications on to Sun systems.
Networking Facilities - how to get the best from them, February 1989 (date to be notified later) Rutherford Appleton Laboratory
