The Annual Report this year does not cover a complete year going from September 1989 to June 1990. The aim was to bring the Annual Report into line with the first half of the Reporting Year for scientific staff and to avoid the inevitable delay caused by summer holidays. In the event, some contributions were very slow appearing. In consequence, the holiday season arrived before the completion.
The main purpose of the report continues to be for internal use in the Department indicating what has been achieved and by whom. It also gives new members of the Department some idea of the work programme in total and its objectives. The set of Informatics Handouts together with the latest Annual Report provides a good overview of the last year's activities.
The Divisional structure has not changed throughout the period but the Division Heads have:
The recruitment of K G Jeffery to fill the Systems Engineering post means that 1990-91 may be a year when all four Grade 6 posts are filled for the first time - let us hope that no-one falls under a bus or gets lured away!
The major funding in the Department still comes from the Computing Facilities Committee of the Engineering Board. The major function for CFC has been the development of the Engineering Applications Support Environment (EASE) with its emphasis on Awareness. This year the events organised have grown considerably with the annual EASE Conference at Manchester being a great success. The Newsletter has increased its circulation significantly. Through the year, a large number of visits to HEIs was made and this has given us a much greater insight into what the users required and raised the level of awareness of EASE. The big challenge for the coming year will be in the production of the first version of the EASE environment.
The joint SERC-DTI collaborative research programme eventually decided to provide EASE-like support for the researchers in that programme. While part of the effort in 1989-90 was finalising the Alvey programme, work did start on activities relevant to the new programme. Hopefully, the merit of this work will be recognised and the funding will continue for the period of the programme.
This year has seen the level of activity in the research area increasing significantly, particularly from the European programmes. A substantial level of funding in the Department now comes from external sources. The quality of the Departments research strength in IT is being recognised. One significant event was the invitation by ERCIM, a consortium of European research laboratories, for RAL to join as the prominent UK IT research laboratory. This will strengthen our ability to do joint research in Europe.
This year saw the retirement of two long-standing members of Informatics and RAL's engineering design staff Bryan Colyer and Jim Diserens (actually August 1990). Both played a prominent role in pioneering the use of interactive computing in engineering design with a number of major achievements. Some compensation for this loss was the recognition of David Duce's work in formal specification by the award of an Individual Merit Grade 6 promotion.
Appendix A gives the September 1989 Organogram for the Department, while Appendix B gives the Organogram at the end of June 1990.
Systems Engineering Division consists of two groups, Knowledge Engineering and Software Engineering. Both groups are small in RAL terms. Each has a dual role of direct participation in the Department's support of the Engineering Board's programmes and a research programme which generates knowledge, skill and technology, to be fed, indirectly, into Engineering Board support and other programmes in which RAL is involved.
For the Knowledge Engineering Group, the year has seen considerable technical progress on the ESPRIT projects and industrially sponsored projects which were agreed in the previous year.
The Group devoted considerable energy to the preparation of proposals in response to the Call from the SERC/DTI IEATP Programme. Whilst technical approval was given to several projects, only one, Gateway, has come to fruition as an IEATP funded project. A second proposal, whilst not funded as an IEATP project, has resulted in a direct collaboration between STC, the University of Buckingham (through a CASE Student-ship award) and RAL, in the area of Strategic Investment Decisions in IT. Effort has also been devoted to seeking further ESPRIT funding. The effort of fund-raising is very considerable, and the KE Group is establishing a good track-record of success, for which the reward is an exciting and stimulating programme of work.
March 1990 saw the completion of the Alvey IPSE 2.5 Project in which RAL was a partner. The project came to successful conclusion in terms of RAL's technical work which was undertaken jointly with the University of Manchester. This project provided a sizeable fraction of SE Group's resources and considerable effort was devoted during the year to finding continuation' funding for the lines of work started in the project. Proposals were prepared for SERC, the IEATP Programme and ESPRIT. This time-consuming work has met with some success, and new programmes will be starting later in 1990.
The equational reasoning work, started by Dr Jeremy Dick, is continuing through an IEATP project on Verification Requirements for LOTOS Specifications. Part of this project is also using specification of graphics systems as a case study.
Both Groups have contributed to the EASE programme in various ways, including participating in the HEI Visits Programme; SE Group have been heavily involved in the Environment and other working groups.
The main perturbation in the Division was the departure of the Division Head, Dr R W Witty, on five year secondment from SERC as Head of the Systems Engineering Centre at ISPRA. David Duce ran the Division on a temporary basis from 6 November 1989 to the end of the reporting year. On 29 June 1990, Dr K G Jeffery from Central Computing Department was appointed to the permanent post of Head of Systems Engineering Division and David Duce received an Individual Merit promotion to Grade 6, effective from 1 July 1990.
In spite of the difficulties, the Division has produced very high quality technical work during the year and has strengthened a number of important collaborative links with academic and industrial research groups, as well as contributing to the EASE programme. The challenge for the future is to improve the technology transfer from the research programmes in the Division to the EASE programme.
MMI2 is a five year research project drawing on 60 man years of effort that started in January 1989 with funding from the CEC under the ESPRIT initiative. The consortium undertaking the project consists of 2 software companies, 2 universities and 3 research laboratories; that is BIM (Belgium) as prime contractor, with Intelligent Software Solutions (Spain), University of Leeds (UK), Ecole des Mines de Saint-Etienne (France), SERC/RAL (UK), ADR/CRISS (France), INRIA (France).
The objective of the project is to develop a highly interactive multi-modal interface for human-machine interaction with knowledge based systems. More specifically, the project aims to build a human-machine interface which will:
The interface will provide simultaneous interactive use of modes suitable to support the developed skills of professional users, and natural communication modes well suited for naive users. These modes include: Natural Language typed at the keyboard (English, French and Spanish), Graphics with Direct Manipulation, Mouse Gesture, and Command Language.
The interface will initially be developed to accommodate a Prolog based Knowledge based system for local and wide area computer network design. This provides the requirement for intelligent dialogue, different classes of users, and the integration of multiple modes of representation and communication. The interface will be designed to be portable across a range of potential applications of knowledge based systems. Tools for the rapid adaptation of the interface will be developed in the later stages of the project.
The overall architecture of the project can be viewed as a three layer structure with the different interface modes represented by modules (or experts) in the upper layer. The central layer contains six experts to manage dialogue and the lower layer represents the application program. The six experts in the dialogue management layer may be imagined as being placed on the points of a pentagram with one in the centre. The central expert acts as a central switching facility to direct information packets between the other five dialogue management experts; the second communicates with the application; the third stores information about the state of the screen and user interface; the fourth constructs and manages a model of the user; the fifth manages the dialogue context; and the sixth manages the semantic information required to communicate between the user and the interface.
Whereas the first stages of the project reported last year mainly covered initial design, this year's work has mainly involved prototype implementations of the major components of the system to allow studies of their interaction to be undertaken. The Knowledge Engineering group at RAL have been involved this year in four tasks within the project which will be described individually.
The CMR is the language used to pass information between the experts in the overall architecture. It uses a typed first order logic that uses promiscuous rectification of events and relations, augmented by communication forces. Expressions in this language are then put into packets along with details of the mode (Graphics, English etc.) from which the information is derived, the time the packet was constructed and other information which may be relevant to later processing.
The formal specification of this language was produced in 1989, although the details of the predicates which can be included in it, the cases of those predicates and the set of communication forces which are admissible have all been further developed since then, as they will continue to be throughout the project.
Following the progress made last year in establishing an architecture for the user modelling expert, work this year has implemented that architecture and incorporated stereotypical knowledge about users derived from experimental studies.
The architecture proposed will acquire information about the user, store that information and provide it to other experts when it is required. The acquisition takes place by three methods: i) stereotypical knowledge of users is derived from experimental studies and entered into the user model prior to use; ii) when the system is running the information passing from user to system and system to user is overheard by the user model and user's beliefs and preferences are extracted from this; iii) in the last resort, direct questions (such as 'what is your name?') are put to the user.
The information is stored in a network of stereotypes. The stereotypes at the top of the network represent general beliefs, goals, preferences and plans of most users. The stereotypes at the leaves of the network represent the individual users themselves. Between these are stereotypes of increasing specificity. If a belief is not explicitly represented in a user's own stereotype model, then the stereotypes up to the top of the network are successively searched to find that belief. If the belief, or a contradiction is not found, it is assumed that the user does not hold it since the model of the world is closed and that which is not explicitly stated as true is assumed to be false.
As beliefs are added to a user's model they are tested against beliefs in the stereotypes from which that user inherits information. If they contradict any existing beliefs the stereotype which contains the contradictory stereotype is removed from the inheritable set. After information has been added to a user's model, and all contradictions removed, a list of rules is consulted to assess if the new set of information about the user implies that the user model should inherit information from stereotypes which it does not; and adds them if appropriate. This process maintains the truth of the information about users by implementing a non-monotonic logic; although at a cost for run time efficiency.
The beliefs, goals, preferences of user's, which are stored in user models, differentiate between knowledge, belief, and awareness. Logical closure under belief is not included since users clearly do not believe all the implications of any belief and often maintain not only beliefs whose implications are contradictory, but also explicitly contradictory beliefs.
The experiments to derive the stereotypical information about users and the content of the rules to assign stereotypes to user models have been conducted by our French partners at INRIA. The data from these experiments have been analysed and incorporated into the implemented user model.
The user modelling expert is now being tested in conjunction with other parts of the overall system.
Much of this year has been spent implementing the architecture for the graphics mode to allow CMR packets produced by the system which contain numeric information to be displayed to the user as histograms, pie charts and line graphs. Users can then alter or query values in these representations through direct manipulation, which will then result in CMR packets being sent to the rest of the system.
At present the tools for producing graphics are being used in conjunction with other parts of the system to explore the structure of multi-modal dialogues.
It is planned to further develop this graphical interface in two ways. Firstly, to investigate the semantics underlying the manipulation act of changing or referring to components of graphical displays. For example, if a user types the English sentence Tell me more about <select> (where <select> is an act of selecting the title of a graph with a mouse) does the user wish to know about graph titles in general, this one in particular, or is the user using the title of the graph as a symbol to refer to the body of information displayed on the graph? Secondly, to develop a knowledge base which will automatically design layouts for graphs using the best principles from ergonomic and graphic design research.
Leeds University have developed and implemented an algorithm for identifying design gestures of the class normally used to mark up manuscripts for editing when they are drawn on the screen using a mouse. At RAL we have developed the semantics of these design gestures so that the actions can be coded in CMR and transmitted to the rest of the system.
As the CMR is used to represent the meaning of user and system actions at the interface, so there must be a language to represent the knowledge about the domain within the application KBS. This representation must not only be sufficient to support the reasoning about the domain which the KBS itself will perform, but also to allow the interface to interact with it in order to support co-operative dialogue and advanced explanation. A specification has been produced of a frame based representation language for the computer network being designed by the user and KBS, with the major inference structure being one using a skeletal design within which modules are individually designed in detail.
The Knowledge Engineering Group is undertaking a project for a consortium of water companies coordinated by the Water Research Centre, Swindon. This project is a development project which is intended for actual use by the Water Industry and provides the Knowledge Engineering Group with valuable experience of development in a practical engineering domain such as those that might be found in EASE. The aim is to develop a pilot knowledge-based system for assisting in the management of water supply systems. The project is divided into four phases: the Definition Phase, in which visits to sites of all the members of the consortium were made, common problems suited to a knowledge-based solution were identified and a pilot site was selected; the Prototyping Phase, in which knowledge acquisition and system development take place at the pilot site leading to a version of the system complete in coverage though not in depth; the Pilot Phase, in which the prototype is enhanced to yield a genuinely usable system at the pilot site; and the Platform Phase, in which the site-independent knowledge is abstracted and a platform constructed on which the other members of the consortium will be able to build their own applications.
The Definition Phase was completed in December 1989. Two or more Knowledge Engineering staff had visited sites from each of the nine consortium members and had conducted interviews with several members of staff. The interviews were a preliminary knowledge acquisition aimed at scoping the project: searching for common problems and opportunities for a knowledge-based system. A report was written and accepted by the consortium based on the findings of this phase. It identified three general classes of application which were found at the sites visited and which could benefit from a knowledge-based system (KBS) solution, namely assistance in dealing with non-routine abnormal conditions on the supply system, risk assessment and resource management. Furthermore, the Southern Water site at Otterbourne was chosen as the pilot site on account of the high quality of its experts and the typicality of its supply system and working methods.
The Prototyping Phase is now under way. A period of intensive knowledge acquisition lasting some three months took place with two principal experts at Otterbourne. Again the interviews were conducted by two or three Knowledge Engineering Group staff, recording what was said and afterwards typing up transcripts of the sessions. The transcripts were later analysed according to the precepts of systematic knowledge acquisition to identify objects - both material and conceptual - in the experts' view of the domain, the relations between objects and the kind of strategies and problem-solving methods used. The transcripts were used to create an intermediate representation of the domain in readiness for the system design and coding which followed. At this stage the knowledge acquisition was driven by a need to develop an understanding of the domain itself - how a water supply system operates - and of the place of the three applications within it.
When a suitable point had been reached, that is, when it was felt that enough knowledge had been acquired to allow a definition of the functionality of the very first version of the prototype, the building of the system began. An AI toolkit, GoldWorks II, had been chosen for the purpose. It is being run on a portable Compaq personal computer, which has the advantage that the software can be demonstrated to the experts in their own places of work. The design method adopted strikes a balance between unstructured rapid prototyping and rigorous top-down requirements-based design, in that it takes advantage of the debugging features and rapid development cycle of GoldWorks, but by means of a series of Design Notes assumptions have been made explicit and the design modu1arised. The system will comprise four components: a model of the water supply system, a situation assessor for reporting on the consequences of given courses of action, an option generator for suggesting actions to the user, and a graphical user interface.
The first working version of the prototype has been taken to Otterbourne for feedback from the experts, who received it favourab1y. It has also been demonstrated to the full consortium. The tasks ahead are to enhance and extend the prototype in preparation for a period of trials and feedback with the consortium and the Southern Water experts in August.
The Knowledge Engineering Group has been engaged to undertake consultancy on a study being carried out at British Telecom's Research Laboratories at Martlesham. The study is on the theory of decision making, and the Group's experience in knowledge representation and knowledge acquisition both in theory and practice are valuable. A number of documents produced by the study team have been reviewed, and some BT review meetings attended. The study is comprehensive in scope, dealing with both single-person and group decisions and normative and descriptive approaches. There is some possibility of a further role for the Group if the findings of the study lead to an attempted implementation of a KBS, as was at one time hoped.
The project, a collaboration between the University of Buckingham, RAL and STC, will produce a knowledge based tool to aid senior management in making and controlling strategic IT investments. Currently we are in the scoping phase during which we are acquiring knowledge from senior managers in STC to establish the concepts used and tasks performed in the domain. This will form the basis for a high level characterisation of the system and the subsequent detailed knowledge acquisition and prototyping tasks. The project forms the basis for JAD's PhD work which is funded by a CASE studentship at the University of Buckingham, with STC as the industrial partner.
The Gateway project is a collaborative IEATP project (1751). The partners are Logica Cambridge, Integral Solutions Ltd. and the Rutherford Appleton Laboratory. The project started in April 1990. It aims to meet the need for better defined and more standard measures for evaluating the KBS elements of system performance; for standard and systematic guidelines for assessing the suitability of a knowledge-based approach and quantifying its feasibility; and for sound comparative evidence from evaluation to support the case for KBS in preference to some other approach. In this way the project is highly relevant to EASE because it will promote an increase in confidence towards KBS projects and will encourage uptake of KBS by allowing meaningful comparisons to other approaches.
There are three areas in which evaluation is appropriate and which are to be addressed by the project: system performance (software and hardware considerations), usability (interfaces, maintainability, user motivation, etc.) and organisational effectiveness (relation of the KBS to the milieu in which it operates).
Specifically, the objectives are as follows:
The project is drawing on evaluative measures for conventional systems where appropriate and modifying them for application to KBS. It will also develop measures for KBS-specific aspects. Software tools will not be developed from scratch but will be employed and evaluated where possible and minor adaptations will be made where appropriate. It is intended that comparative evaluation will lead to some absolute measures, thereby enhancing the predictive power of the methods, which is an important goal because of the need for prediction in the early stages of a system life-cycle when it is necessary to assess the potential for KBS involvement and to present evidence for its effectiveness.
The Work Packages into which the project is divided are as follows: Data Collection and Analysis, Synthesis of Metrics, Evaluation of Metrics, Further Synthesis, Further Evaluation, and Integration. Thus there are two cycles of synthesis and evaluation, which allows for feedback from the results of evaluation on real projects which it is hoped to be able to carry out. At present Work Package 1 is in progress: a wide-ranging survey of metrics and related topics in conventional software and knowledge-based systems has been carried out, and a conceptual framework is being developed in which metrics may be organised so that their utility in context will be illuminated and gaps in coverage revealed, thereby setting the scene for the synthesis of metrics which follows in the second Work Package.
Specifically:
During this year, a difficult year for formal methods groups due to lack of funding, the Software Engineering Group has remained at about the same overall numbers. With the IPSE 2.5 project completing, we have had the LOTOS and ZIP projects starting and have been active in submitting further research proposals that are about to come to fruition later in 1990. This should allow us to slightly expand our numbers.
The main changes this year have been that Mike Spivey has been away in the States on sabbatical and that Ian, a "young scientist", has joined us for 6 months from June 1990 to work on a browser for PCTE's OMS.
In August 1989 DAD received his chair from the University of East Anglia. BR received his Professional Review for C. Eng status; the interview panel have subsequently recommended to the BCS that he be awarded this. In June 1990 DAD and SKR formally received their C. Eng.
Within the group, software quality is approached from the formal methods angle. Two overall objectives drive the work:
DAD has continued work on graphics standards. DAD and F R A Hopgood (FRAH) attended the ISO/IEC JTC1/SC24 meeting and the Computer Graphics Reference Model meeting. The meeting was not overly productive, but progress was made on the Reference Model at a subsequent meeting in January 1990. In collaboration with the LOTOS project and Graham Reynolds at the University of East Anglia, work has started to explore the extent to which the structure of the Reference Model provides a structure for a formal specification of a graphics system. This exercise has already produced some useful insights into the Reference Model which form a part of the UK comments on the Reference Model for the next meeting in July 1990.
DAD and FRAH resigned as document editors for the GKS Review at the October 1989 meeting following a decision to take the existing ISO 7942 document as baseline rather than a UK document which in their view provided a better structure from which to start the Review. Since that meeting DAD and FRAH have contributed to UK efforts to pursue some of the key ideas in the UK input to the GKS Revision, in the area of the relationship between GKS and the Computer Graphics Metafile and user extensible output primitives. This work has culminated in the production of two new documents, known as GKS-P and GKS-Q which form part of the UK input to the GKS Review meeting in July 1990.
DAD and FRAH gave a 2-day tutorial on the Programmer's Hierarchical Graphics System (PHIGS), published as an ISO standard in 1989, at the Yugraph 90 conference together with invited lectures on Formal Methods in Computer Graphics and Super Workstations respectively. Prior to the Yugraph conference, the PHIGS tutorial was given at RAL as a part of the EASE Awareness Programme.
IPSE 2.5 was a major Alvey project to research and develop an Integrated Project Support Environment based on advanced distributed systems and man-machine interfaces, and incorporating support for both the practice and organisation of design tasks. A major objective was the development of an integrated framework for supporting the use of formal methods in software development, including formal specification and theorem-proving techniques.
The project began in October 1985 with three initial collaborators (ICL, STC and Manchester University). RAL joined the project in July 1986 and three additional industrial partners subsequently joined the collaboration. The industrial component of the project terminated at the end of 1989, the Manchester University{RAL component at the end of March 1990.
At RAL, Juan Bicarregui (JCB) and Brian Ritchie (BR) were engaged in the project, both spending a fair proportion of their time at Manchester. DAD was the RAL representative on the Project Review Board.
The Manchester/RAL part of the IPSE 2.5 project (Theme C) was concerned with the support of formal methods of software development and in particular with aiding formal reasoning itself. The intention was to build tools which enable a user to construct proofs at the workstation; modern workstations such as the SUN3 and SUN4 make it possible to design proof assistants which are much more usable than earlier tools developed around glass teletype interfaces.
Manchester concentrated upon the specification and implementation of a generic proof assistant; RAL concentrated upon implementation of a VDM Support Tool for both specifications and designs using BSI VDM. The resulting system was named Mural. Mural provides an environment where specifications (and the refinement of one specification by another) in the VDM support tool can be "translated" into theories in the generic proof assistant, where the proof obligations that justify a specification or design step can be discharged. Originally, the VDM support tool was intended to be "minimal", the main aim being to provide an existence proof that the generic proof assistant could be instantiated to support particular formal methods. However, with the withdrawal of the IPSE 2.5 theme responsible for generic formal methods support, the VDM support tool gained greater significance.
BR and JCB's work in the last months of the project was divided between developing an improved interface for constructing VDM specifications and developments, and implementing the translation mechanism for the generation of appropriate theories in the proof assistant from such developments, including proof obligations. Though the final support tool is not a fully-fledged support system for VDM (in particular, it only permits a restricted form of specification refinement), far more was achieved than was originally envisaged.
A first draft of a Book of Mural has been submitted to a potential publisher. This documents the system and the designers' experiences in implementing in Smalltalk from formal specifications. BR and JCB have contributed two chapters on the VDM support tool. Further work on this is anticipated.
Further, it is anticipated that Mural will be of use in RAL's contribution to the IED Raise User Trials project which is expected to start soon.
This work was initially funded by SERC's Computing Science Sub Committee. The ability to reason with equations (for example, to deduce conclusions from a set of equations), is important in a number of areas, such as specification of abstract data types and their validation; program transformation; synthesis of programs; program optimisation and solving equations. Dr Jeremy Dick had, over a number of years, developed a rich theoretical framework (at RAL) for equational reasoning and had developed a practical tool, ERIL (Equational Reasoning - an Interactive Laboratory), based on this framework. ERIL is a useful tool for experimenting with specification and prototyping of abstract data types, and is used by several research groups in the UK and abroad. ERIL is based on the use of rewrite rules for computing and reasoning with equations. The project formally finished last September and BMM produced the final report for SERC.
Building on work undertaken in his MSc, BMM carried out a series of experiments in using ERIL in various configurations for theorem proving, in September 1989. The results of this will be appearing in a lab report.
As part of the IPSE 2.5 and LOTOS projects, JCB and BMM designed a new higher-order matching algorithm for use within the Mural theorem proving system, using VDM specification, and rapid prototyping in Standard ML. This proved to give a significant improvement in the performance of the system. An account of the development of the algorithm has been written up and will be presented at the UNIF'90 workshop at Leeds in July 1990.
CMPR, DAD, SKR, BMM and JCB are involved in the SERC IED project Verification tools for LOTOS Specifications with RHBNC and Glasgow which started in October 1989. This continues the work undertaken in the ERIL programme.
This is a three year project funded by the IEATP. The consortium consists of Royal Holloway and Bedford New College (RHBNC), the University of Glasgow and the Rutherford Appleton Laboratory (SEG) and is "unc1ed" by British Telecom (BT).
Formal Methods of specification and verification are a technology which must become more widely used by UK companies if they wish to remain competitive in the long term in areas where they are becoming the industry standard. One such area is the formal specification of standards for open distributed systems, and in particular those related to Open Systems Interconnection (OSI) computer network architecture, for which the specification language LOTOS (Language of Temporal Ordering Specification) was adopted as an international standard in 1988 (IS8807).
LOTOS is based on the concept of specifying a system in terms of observable behaviour. In LOTOS events are used to denote the occurrence of something which the specification makes assertions about. LOTOS can express ordering constraints on events and through their structure the communication and change of information within a system. The ordering and structure of events are captured in two separate components of the LOTOS language. The first of these, the process calculus, is based on process algebraic methods (Milner) and the semantics are based on Milner's CCS (system of inference rules and labelled transitions of processes) and Hoare's CSP (operators for process composition). The second component, based on the data typing language ACT ONE, uses equational specification with initial algebra semantics.
The industrial use of LOTOS has focused on developing formal descriptions of OSI standards. Complete formal descriptions of several OSI standards now exist and are being progressed within ISO as technical reports. It is ISO policy that these texts will become definitive standards. The project uncle, British Telecom, is heavily involved in this work, and Glasgow University is already involved in LOTOS work involving the translation of ASN.1 into LOTOS.
If full benefit is to be obtained from the use of this formal specification language it is essential to be able to reason about such specifications and the implementations derived from them, in order to determine that the specification has the required properties, or verification requirements.
BT is already working on the verification of LOTOS specifications, using existing theorem provers. However, while research has been done on verification for different parts of the language, it has not yet been incorporated into an integrated whole. Term rewriting and equational reasoning are clear candidates for this work, as they are already widely accepted as a verification technology for the major constituents of the language, algebraic specifications and process algebras.
The equational reasoning system ERIL has been developed at RAL, with significant technical research input from RHBNC, and is particularly suited for experimental use as it is easily user configurable to allow experimentation with a wide variety of strategies. While it is felt that as it stands ERIL could be of significant use in discharging the verification requirements of LOTOS, the project members also believe that a long term research project such as this will allow significant further technical developments.
Therefore, the objectives of the project are
The work in (2), undertaken by RHBNC, will also indicate directions in which the ERIL tool has to be enhanced to serve as a test-bed for these ideas. The development of the ERIL system in the light of this falls to RAL. The developed system will be assessed using examples arising from the case studies in the final year of the project.
The result of the project will be
BMM has been designing and implementing a basic set of general tools for use with equational reasoning. This is being undertaken in the functional language Standard ML. Initial experiments in producing reasoning tools and interface tools were undertaken and a Knuth-Bendix Completion algorithm was produced. This has now been developed further with the use of the ML module system to produce an initial basis for a set of tools for equational reasoning. In addition an ERIL like interface has been placed on top of the tools to give a clean presentation of the tools to the user.
An initial design document was written and presented to the project meeting of the 13th March 1990.
Worthy of special mention is the development of an incremental mixfix parsing algorithm which allows the user to add and delete new function symbols to the equational reasoning system as he or she chooses, the function symbols having a free mixfix format. The parser for terms is then adapted dynamically to recognise the new symbol.
As a case study in the use of the LOTOS specification language, work has been undertaken in the formal specification of Computer Graphics Standards using the LOTOS language. LOTOS has been used largely for communication protocols, so this is a new area of use for the language. It is intended that properties of the graphics standards can be demonstrated by analysing the specifications. All this work is still continuing.
Building on previous work of DAD and E. Fielding, BMM reformulated a description of GKS in OBJ into a ACT-ONE specification. This proved a straightforward task, but the specification was not strong enough to allow proof of properties.
DAD and BMM are describing the Computer Graphics Reference Model (CGRM) in LOTOS. An initial highly abstract description of the model has been produced, which shows some of the difficulties of describing a state-based model in a algebraic framework. This is now being modified, and the functionality of GKS output is being fitted into the model.
CMPR has been studying GKS input mechanisms, structuring the Logical Input Devices and Input Modes into LOTOS. He has also been working with James Purvis at Brunel University on GKS input/output expression in LOTOS.
ZIP is a three year project employing research and development effort from the membership of British aerospace (Military Aircraft) Ltd (Lead), BP International Ltd, IBM UK laboratories Ltd, Logica Cambridge Ltd, Praxis Systems P1c, Rutherford Appleton Laboratory and Oxford University Programming Research Group. Work began on the first of February 1990. The RAL contribution involves SEG's (PRG) Atlas Fellow Jim Woodcock.
The work programme schedules activities for:
The standardisation of the Z notation, its mathematical toolkit, syntax [PRG] and machine processable representations [Praxis], taking account of further notation development resulting from research into refinement and concurrency.
Z is a mathematical notation for expressing the specification of a computer system or program, which can also be used in the design process. A specification is the means of communication between a client (user) and a supplier (implementer); additionally it is a mode of communication between team members. To succeed it must be clear and unambiguous, which existing informal methods of specification are not. Formal notations have been developed in response to this need.
In the last few years work has proceeded apace, both at the PRG and in industry, to agree a common abstract syntax for Z. With the publication of the Z reference manual (Mike Spivey, 1989), there is a basic reference for Z - the essential core of Z is now stable. Further research needs to be pursued, especially in the area of the specification of concurrent systems, and this could result in extensions to the notation. This research work forms the core of RAL's contribution to the ZIP project (through Jim Woodcock).
Currently there is no public standard for Z. The Z notation is now sufficiently mature for the development of a standard, and has gone through the early development stages when the language was changing rapidly but has not yet suffered disparities. Such a standard needs to be subject to control so that changes are carefully documented and published in a regulated manner. The body to agree such a Z standard should include representatives from industry and academia. In the past, the extent to which involved parties had entrenched positions militated against agreement on programming language standards. That is, the standards were always too late! In the Z world, the time is ripe to agree a standard before such inflexibility develops.
If Z is to become widely adopted then a fixed point of reference is needed for the notation. Other, older formal notations have suffered from a divergence in representation and use and consequently have become unwieldy, and exchange of documents has become difficult. Most people use some form of text processing system, and it must be possible to send machine-readable formal specifications from one organisation to another. It then becomes vital the writer and reader use the same language - a standard concrete syntax for Z is needed. One additional requirement is for textual representations which can be substituted for Z symbols on devices which are incapable of displaying these symbols.
The application of Z in industry to provide case-study examples and heuristics for the methodical use of Z in a fully formal system and software development process. [Logica Cambridge]
Although Z has a well defined notation, like many other formal methods there is little guidance as to the manner in which it should be used. This deficiency has been highlighted by the Alvey Formal Methods Advisory Group. Z needs to have a developed method if it is to become widely used.
At present there is a great interest in Z, stimulated in part by DEF-STAN-0055, as is witnessed by the increasing numbers attending courses in Z (Praxis, Logica, PRG). Unfortunately, these converts to Z experience difficulty in applying the notation because they cannot be sure that their language use is correct. Moreover, they find writing Z arduous. One reason for this is that there is no available guidance on what constitutes a good Z specification, nor on how to approach specification problems.
The development of prototype tools adopting the standards of (12) and supporting the methods of (2). [BAe]
The state of the art in tools supporting software development using Z is represented by a number of text formatters, editors, type-checkers, and prototype proof assistants. These tools have been developed by individuals and organisations to varying degrees of quality using various machine representations of the formal notation. These include the FORSITE tools (RACAL, System Designers, PRG, Surrey University, RAL) FUZZ (J.M. Spivey, RAL), BAe ZED, instance of Genesis (1ST Cambridge), IBM Z tools, ZEBRA Type-checking Environment (Bernard Suffrin, PRG), use of the B tool (Jean-Raymond Abrial with PRG and BP), Logica Z Editor, Imperial College Z Editor and RSRE Z Editor and Type-checker.
The basic and applied research to support these standards, methods and tools. This to include work on concurrency and refinement [RAL, BP].
The major products of these research and development activities will serve to promote the unification initiative and enhance the state-of-the-art.
The Esprit Elusive Office (ELO) project aims to investigate and provide state-of-the-art solutions to the software and hardware requirements for supporting distributed office working (or tele-working). The prime contractor is Empirica (FRG); together with ISI (FRG), they are investigating requirements. SEL (FRG) are providing hardware and communications expertise. RAL's contribution to the project is largely in:the area of software support (together with Computer Lern Systeme (FRG) and Realace (EIRE)).
Since April, BR has been assisting Design Division's contribution to ELO, drawing on his previous experience in designing within an object-oriented environment to develop the ELO Integration Framework. It is planned to use existing toolkits and class hierarchies as far as possible to provide a starting point, though further work is required to provide the level of integration required by the computer-aided learning support planned by CLS, and the security services planned by Realace.
BR has taken part in a number of ELO meetings, both with software and requirements partners, and has helped shape the requirements upon the integration framework, and investigated possible approaches to solutions. This has included the investigation of some existing C++ class hierarchies.
The current Engineering Board supported research community tends to be split along Committee lines. This does not take into account computing solutions that often straddle several communities' application areas. Therefore, CFC is taking an active role in facilitating cross-fertilisation between Committees, disciplines and industry in the computing area and, as part of this activity, the Engineering Applications Support Environment (EASE) programme has been initiated with the mission statement:
To stimulate and encourage engineers to use appropriate, state-of-the-art software and hardware to enhance their research
The programme itself involves many activities. SEG's participation in the programme is outlined in the sections below.
Specifically, CFC wishes to encourage engineers to migrate to state-of-the-art integrated project support environments (IPSEs). These environments were assessed last year (by SKR and JRK, as part of the EASE programme) as not being viable until the mid 1990s and so a working party, involving SKR, has been exploring the means whereby a software environment that would provide a transition path to this technology could be devised. Part of this activity is the continued tracking of IPSE technology.
As part of this tracking process, JRK has undertaken an assessment of the Portable Common Tools Environment (PCTE) as a Tool Support Interface (TSI), using an Emeraude implementation of PCTE installed on a SUN 3 Workstation. JRK and SKR have made contact with potential suppliers of IPSE technology (Softlab and IPSYS) and have been involved in several visits / meetings to discuss our requirements for the Engineering Applications Support Environment (EASE) . A survey document summarising these various Environments activities has been produced.
SEG have recently (June 1990) taken on a Trainee Programmer, IRJ, to work on a 6 month project, supervised by JRK. IRJ's task is to write (in C) a user-friendly browser for the PCTE object store. The motivation for building this tool arose directly out of JRK's experiences during his assessment of PCTE.
A working party was set up in March 1990 to investigate the feasibility of an EASE Information Directory (EID), and has just submitted an Interim Report to IDHM. JRK, together with colleagues Deborah Thomas (DD), Subodh Chandra (DCSD) and Bob Maybury (CMD), have been tackling two main areas of work so far. The first has been to establish what the HEI's requirements for an information service are: the second has been to find out what information services are already in existence and to determine the extent to which they satisfy those requirements.
Part of the EASE programme involves a survey of current IT practises and perceived future IT needs within the EB community. JRK, DAD and SKR have participated in this activity that involves an initial visit to a significant number of the HEIs in the UK, followed up by continued contact (to monitor the effect of the EASE programme).
During the visit guided interviews (using a questionnaire) are undertaken with members of the community and the findings thereby recorded. For example, JRK and Neil Calton (from DCSD) carried out the initial visits to Bradford University, Huddersfield Polytechnic, Sheffield Polytechnic and Sheffield University.
Using the Questionnaires from 10 such HEI visits (ie. about 100 in all), JRK made a very preliminary assessment of the findings of the HEI survey, in terms both of current practices and of requests for EASE provision. This formed the basis for short presentations which JRK made during the visits to Softlab and IPSYS. More detailed analysis is being undertaken within the EASE Environment and Information Directory activities.
BR remains active on the BSI VDM committee, although his other responsibilities have overshadowed his involvement somewhat. His contributions over the past year have largely involved maintaining the source of the draft standard document. ISO has voted to accept the development of the BSI work as an ISO standard, so it is anticipated that this work will continue.
BR is the Monitoring Officer for the IEATP project 1563, Foundations and Tools for Formal Verification (now named FST). The lead organisation is ICL Defence Systems Ltd. (project manager R. B. Jones), the other partners being Program Validation Ltd., University of Kent and Cambridge University. The project started in January, and will run for three years. The monitoring effort involves assisting in the production of a Project Plan, attending project meetings, and drafting quarterly reports on the technical and financial progress of the project.
SKR is the monitoring officer for three other IEATP projects (1150, 1253 and 1266) while DAD is monitoring the AniMate project.
JCB and BR are contributing towards an evaluation of the SL-GMS UIMS for the EASE programme. It is planned to determine the degree to which a high-quality interface can be constructed for an existing (preferably Fortran) program with as little alteration of the program structure as possible. A copy of SL-GMS has been obtained and installed; unfortunately, it lacks the Fortran bindings that are needed for this part of the evaluation. At the time of writing, SL's UK distributors have promised to send suitable bindings.
DAD and Ruth Kidd (DCS Division) have been exploring the applicability of ISO/IEC document interchange standards to the problem of exchanging technical documentation and other documents within the HEI and Research Council's establishments servers by the SERC/Computer Board Advisory Group on Computer Graphics (AGOCG). A joint paper was submitted to the AGOCG SGML Workshop in march 1990 on experiences with using the Standard Generalised Markup Language (SGML) to describe parts of the RAL GKS User Guide.
The Workshop recommended that some experiments be carried out using SGML and AGOCG subsequently endorsed these. DAD, BMM and RMK have subsequently devised a workplan for this and are in the process of mounting the German Research Network's (DFN) DAPHNE software at RAL. DAPHNE is a translation tool for converting documents described with SGML Markup to the markup of a target formatting system such as troff or LaTeX. This will be used as the basis for the experiments. So far, the DAPHNE system has been acquired, but the first version received was mis-configured and a replacement has only just been received.
In March, the group gave a one day tutorial to a group of trainee Young Scientists entitled An Introduction to Formal Methods. The course covered the role of formal methods in software engineering, the fundamental concepts of specification and refinement, and an overview of some topics of current research in formal methods. It is hoped that the material from this course can be reused at some stage, perhaps through publication.
BMM continues to distribute the Rutherford Standard ML / Cambridge LCF system, although requests for this system are declining. In addition BMM also supports the ERIL equational reasoning system.
DAD contributes to the Department's involvement in the ESPRIT II project, ARGOSI. The project concerns the integration of graphics and networking standards. DAD's main contributions have been to a workpackage concerned with a classification scheme for applications in terms of graphics and networking requirements, and to some of the management aspects of the project.
DAD is Treasurer of the European Association for Computer Graphics. During the year he was Secretary of the Organising Committee for a joint Eurographics/ESPRIT Workshop on User Interface Management Systems and Environments, held in June 1990 and is also co-chairman, with Carlo Vandoni at CERN, of the technical programme for the EUROGRAPHICS '90 conference to be held in September 1990.
Brodlie, K.W., Duce, D.A. and Hopgood, F.R.A., Second Generation Computer Graphics Standards, RAL Report RAL-89-097, 1989.
The overall objective of the EASE environment project is to develop a national IPSE for EASE by the mid 1990s. The environment is to be based on de facto standards and designed to stimulate and encourage engineers to use appropriate, state-of-the-art software and hardware to enhance their research. Its use is to be encouraged by being designed so that it can be populated by a set of tools or modules with wide applicability. The software environment and information directory work can be seen as an ongoing commitment on the group throughout 1990/91
SEG research projects should lead towards increasing SEG's strength in
The ongoing ZIP and LOTOS projects will contribute to achieving these objectives whilst the anticipated RAISE User Trials project should broaden the group's area of expertise. We are keen to integrate the equational reasoning and proof assistant work which, although research proposals in 1989/90 attracted alpha ratings, have yet to attract funding.
Design Division consists of three groups - Applications Integration, Systems Interface and User Interface - together with the Transputer Coordination Unit which incorporates the London and South-East Regional Transputer Support Centre.
The objectives of Design Division are:
Specific highlights of the past year, which are detailed in the sections below, are:
The Division expects that its main objectives in the coming year will be:
In addition to these broad targets, each Group and the Transputer Coordination Unit have identified their own specific objectives for 1990-91. These are detailed, together with a review of 1989-90 activities, in the sections that follow.
Staff in post:
The objectives for the group were to contribute to the developments of standards in engineering data exchange, to increase the awareness of these standards in the academic community, to investigate how these standards can be implemented into a database as an aid to applications integration, and to contribute to the Automotive Design Project (ADP).
A number of articles on the work and achievements of the group have been written and published in the Engineering Computing Newsletter. The articles deal with data exchange in general, results from the CAD*I project, a description of Standard for the Exchange of Product Data (STEP), a description of the Express language, the database implementation work and Computer Aided Acquisition and Logistics Support (CALS).
All members of the group participated in the EASE HEI visits.
Staff of the Interfaces Section have been instrumental in setting up a Community Club on Modelling and Management of Engineering Data. The first open meeting of the club was held on 4 June with 25 attendees, and the first meeting of the Executive of the Club will take place on 31 July. After the open meeting, MM was appointed secretary of the Club. The aims of the Club are to assist engineering academic staff in solving their data modelling and data management problems and to provide a focal point where engineers can discuss problems and solutions. Industrial participation in the Club is being encouraged.
Both sections of the group have given support to the ADP which has its Central Support Unit based at the University of Birmingham. A CADDS-4X system has been installed at RAL, and initial familiarisation with the system has taken place. A project has started to investigate the requirements of the ADP grantholders with respect to hardware, software and networking.
CADDS-4X is a system for Computer Aided Design which is widely used in the automotive industry. The RAL system is running on a Sun-4 SPARC workstation.
A project has started to investigate the requirements of the ADP grantholders with respect to hardware, software and networking (JVM, MM).
KML has just started work to provide the ADP with a central database in which to store all the results from the ADP projects. At present this work is in the definition stage and a questionnaire has been distributed to all ADP grant holders to determine what data they will want to store in the database.
DT and JVM attended a number of meetings of the STEP project. The main aim of attending the meetings is to ensure that the group remains abreast of the latest developments in data exchange and data modelling. During the year, JVM was appointed chairman of the STEP working group dealing with the physical file format and STEP implementations.
JVM and DT attended meetings of BSI AMT/4 which is monitoring the progress of STEP. DT also attended meetings of BSI AMT/4/ - /3 and AMT/4/-/6, the panels which are involving industry in the development of the Finite Element Analysis and Materials parts of STEP.
The file reader and parser, which was originally written for the CAD*I project was successfully converted to deal with STEP exchange files. The documentation of the program has been updated. A back-end to the file reader was written which provides an interface to the Nastran Finite Element program (JVM, AJM).
The Express compiler is finished except that it needs to be brought in line with the latest definition of the language. One output module of the compiler is a pretty printer which reads models written in the Express language and produces a properly spaced copy of the input. Another output module which will generate SQL statements from the Express model is in its planning stage (MM).
The software has been distributed to a number of academic and industrial organisations, both in the UK and overseas (see 3.2.7 below).
The STEP implementation committee are just starting to work on an applications programming interface (API) which will provide independence between applications using STEP and the storage method. This will be very useful for applications integration. KPD is working on a trial implementation to test out the feasibility of using an API accessing a relational database. A neutral database for electromagnetic and semiconductor modelling data has been created based on an EXPRESS schema and a mapping from EXPRESS to SQL. Genera1ised subroutines to access the database are in the final stages of development.
RAL retains the Intellectual Property Rights of the software written in the Group. Under licence, 21 organisations have been supplied with the software. These comprise: 10 UK HEIs, 3 overseas HEIs and research institutes, and 8 other organisations as follows:
From April to June, DT spent 50% of her time leading the EASE Information Directory feasibility study. This is a project which is looking into the requirements of the EASE community for a wide range of on-line information. It is investigating what information is already available and will present a proposal to Computing Facilities Technical Advisory Group and Computing Facilities Committee in September for setting up a. service at RAL. The study is being undertaken by a cross-divisional working party, the other members being Bob Maybury (CMD), Subodh Chanda (DCSD) and John Kalmus (SED).
The group submitted three proposals under the ESPRIT programme, of which one was accepted. The project Integrated Design and Analysis Environment for Advanced Magnetic Devices (IDAM) will run for 2 years and requires an effort of 1.5 my per annum. The subject of the grant is the streamlining of data exchange between a number of packages which are required for the design of electromagnetic devices. These packages are designed for specific purposes like geometric design, structural or thermal analysis, or electromagnetic analysis.
Some initial planning work has been done by JVM and DT this year but it is envisaged that most of the work for this project will be taking place in the next two reporting years.
Mike Mead attended the MSc course on Computer Science at Brunel University on day release. He has successfully passed his second year exams.
The main objective for the next year is to make all aspects of the expertise of the group available to the academic community. We intend to disseminate our results and information as widely as possible, both in the fields of data exchange and engineering databases. In relation to this, the Community Club on Modelling and Management of Engineering Data will be supported wherever possible.
All software developed by the group will be made available to the academic engineering community as part of the EASE project.
Effort will be devoted to participation in STEP and similar activities relating to data exchange with the aim of keeping abreast of new developments in this area. Further effort will be given to developing an implementation of an Application Programming Interface to STEP based on a relational database and the results of this work will be disseminated to the UK academic community.
A significant amount of effort will be devoted to support of the Automotive Design Project headed by the University of Birmingham, which involves other academic institutions and automotive industry partners.
Staff in post during the year were:
The group's major interests are in providing tools, techniques and environments for the engineering community. The tools and techniques are directed at support for highly interactive user interface design; the environments provided are aimed at providing productive, integrated software for engineering applications development and execution. To this end, the Group undertakes a range of activities, from research (with a practical flavour), through advanced and conventional development, to support. Each activity is intended to act as a source of ideas and inputs for the others.
TAW has provided consultancy services on X to a number of groups and projects within RAL. TAW and KR, with John Bovey of the University of Kent at Canterbury (UKC), gave a seminar entitled Introduction to X at the EASE90 Conference. TAW also gave a presentation on X to Design Division. Some work has been done in making software expertise available to provide a path to XllR4 from R3.
There have been two main activities here. The first has been a continuing activity with the RAL toolkit WW. Some time has been spent by AVS on minor bug fixes, and on understanding and rationalising the WW code to ease maintenance. Some student work on circular menus has been generalised so that elliptical and semi-elliptical menus are now supported, with a fast find mechanism which enables choices to be made by movements in the horizontal plane only.
Some time has been spent on understanding, if not totally resolving, problems with inverse video and cursors under both X and SunView. CAAG has proved a valuable repository of background information about WW. Some tracking work has also been undertaken - compiling new versions of WW for X11R3, a new version of ten for general use, and generating SPARC versions of Spy for Support to distribute outside RAL.
The second activity has been an assessment of the toolkits available with X, jointly with staff in DCSD, for the EASE programme. These toolkits split broadly into two groups - those that are Xt Intrinsics based, and the rest. Of the rest, most are written in C or C++. TAW drew up the toolkit evaluation strategy and workplan, which then went through a number of iterations following comments within ID. A major activity following that was to design a questionnaire to gather user requirements for these toolkits. This was a fairly successful exercise, and much useful information has been gained; this information has been used to draw up a list of assessment criteria. Some time has also been spent on familiarisation with some of the toolkits on the evaluation list. An X Toolkits Working Party is being established (under the chairmanship of Professor Peter Brown at UKC), to direct the work and eventually report to the Computing Facilities Technical Advisory Group (CFTAG).
One spin-off activity has been that TAW has given a talk on toolkits to an IEE Colloquium in London.
Almost the whole group has been involved in the HEI visits.
TAW is Minute Secretary to the Computing Facilities Committee.
KR was a member of the Engineering Computing Facilities Executive (ECFE) for a while, acted as a monitoring officer for the UKC EMR, and was part of the team that discussed the proposal by Gosman's group at Imperial College of Science and Technology (ICST) for EASE funding. He has also been involved with contributing to the establishment of the heavily-revised EASE Schedules. Both KR and VAB have provided input on occasion to the EASE Environment design.
This project, funded by the Control and Instrumentation Sub-Committee via the Computing and Design Techniques for Control Engineering (CDTCE) Management Committee, is aimed at providing a common environment for control engineers, for both standard package access and control system algorithm development. The basic environment is developed at UMIST's Control Systems Centre under an SERC development contract, with other developments taking place in other HEIs. RAL's role is to provide management and technical support (but without overall project management responsibility), organise beta testing, and so on. Longer-term, user support responsibilities remain at RAL.
A new version of the ECSTASY 1.5 beta-test software arrived from UMIST in late 1989 and this second beta-test got under way. Direct user support was provided by KGD, with DMR as necessary. Like the first beta-test, this was a 3 month evaluation by 4 academic institutions in collaboration with some industrial partners. A number of problems were encountered with installation, but once these were fixed all the sites were able to carry out the test programme. This consisted of tackling a UMIST supplied problem, followed by a problem suggested by the industrial partner. As for the first beta-test, feedback from the sites was collected informally at a workshop held at RAL and more formally by means of a detailed questionnaire that was filled out by both users and installers of the software. There were noticeably fewer bugs reported, most of the comments being about the overall system design, and overall performance of the system.
The work of integrating TSIM into ECSTASY has now been largely completed, with BPKL doing most of the programming, guided by DMR. The issues outstanding all relate to the way in which current TSIM users wish to access the TSIM facilities and the clash between that and the encompassing ECSTASY system. System documentation for the TSIM interface is now being produced.
Towards the end of the period of this report, the beta 2.0 software was received from UMIST, but the significant number of changes since the 1.5 release has made the system somewhat fragile. The problems so caused are currently being investigated. It is intended that the beta test will take place over the summer, with the autumn being used to generate a productised version based on the last PA Set release.
KR attends meetings of the CDTCE Management Committee and liaises at the management level with the CDTCE Coordinator and UMIST.
At a number of points in the year, problems were experienced with the basic IDUS service which affected our ECSTASY work. Particular problems were filestore corruptions on nfs6, and the time taken to obtain and install the FORTRAN 1.3 compiler from Sun, and its corresponding manuals.
The Energy Kernel System (EKS) is an object-oriented programming (OOP) environment for building thermodynamic analysis software. The majority of the funding for EKS comes from the Building Sub-Committee although our work is funded through the EASE programme. Its main objective is to enable research teams currently investigating various aspects of building simulation, especially thermodynamic performance to use software easily and safely from other groups for those parts of the overall appraisal system that are not directly concerned with their research. A second objective was to enable a collection of validated components/sub-systems to be built up, so that the robustness and integrity of appraisal software could be improved.
This leads naturally to OOP and object-oriented databases (OODB). However, as the OOP approach to thermodynamic simulation is still unproven, it is important to show that the use of the EKS and its OOP approach will bring substantial benefits even in the construction of conventional simulation packages.
Phase 1 of the project, the development of the Prototype, is nearing completion. An initial class hierarchy design capable of supporting current state-of-the-art energy models was completed early this year. Three example problems of increasing complexity were selected, modelling site conditions, modelling the thermal processes within as wall and modelling a complete room including a piece of plant. The necessary classes were implemented for the Prototype, a simple model building program developed and the three test models built.
The experience gained building the Prototype has been valuable. The main findings were that, whereas control of the system and in particular control of the time-stepping turned out to be simpler than expected, the design of the information flow between the objects was much more complex than anticipated. The ONTOS database has proved useful, if a bit cumbersome, providing several facilities that simplify the development of the model building facilities as well as acting as a straightforward object database, but care has to be taken not to become locked into this product.
Though the main function of the Prototype was to act as a test bed for the EKS team, feedback has been obtained from other groups on both the design of the class hierarchy and the functionality required by the end users. Based on this feedback, an external evaluation of the Prototype and experiences in building it, a specification for the final EKS Demonstrator will be drawn up.
DMR acts as the project manager, with only minimal involvement by KR. He provides most of the technical input into the project with BPKL gradually building up expertise in the C++ area.
RAL has been providing support for the Environment Committee's IT Applications Initiative. Apart from direct support for the Coordinators of this program, the main task carried out has been the setting up of a "loan pool" of workstations. This involved obtaining eleven Sun SPARCstations and installing them in various departments around the UK. The intention is that these will act as a seed, enabling the exploration of IT tools and techniques by the departments concerned. Basic system administration courses have been run for individuals from the selected departments and a programme of visits by RAL staff will provide IT expertise and advice.
This is a SERC-funded project to investigate ways with which extensibility can be built into a graphical programming environment.
Most of the detailed work is being done by CAAG, with occasional contributions from KR and VAB. During the year, the design of the User Interface construction tool has been refined. The WW toolkit on which the current work is based has been enhanced to provide the few remaining facilities needed to make the extensible tools being constructed useful. These include a text editing widget and various widgets for managing layout and screen menus. Integration of external command streams with graphical input is now supported by generic tools. Some work has gone into arranging for the graphical selection to be seamlessly integrated with both the WW widgets and the host window system.
A new graphical editor has been constructed, along with a graphical tool for connecting UNIX-style commands directly to a graphical environment. This provides a degree of convenience and flexibility not previously available in a graphical context. The construction of this new software technology has paved the way for the UI tool to be re-engineered into a solid tool for general use, which is now well under way. The general aim is to return to the UNIX philosophy by making small tools which do a simple job well, but allowing them to be combined in powerful ways. To this end, the new editor has hooks to allow it to be controlled by another program (which is normally another graphical tool operating on behalf of the user). A number of experimental " tools have been built which demonstrate various ways of making user interfaces configurable. From these tools, a mail tool and a USENET news browser have been built, though they are not developed enough yet to be useful in their own right.
A seminar on Extensible Graphical User Interfaces was presented at the University of Kent and a similar paper has been submitted for the conference The Many Faces of Design later this year.
Progress with facilities built on the WW toolkit have largely overcome the problems last year with the lack of the anticipated Object Oriented toolkit. WW continues to provide a flexible and appropriate environment for User Interface Innovation.
This ESPRIT II project spans a number of groups within the Department, covering contributions from UIG in DD (TC, MJP) and Parallel Processing Group in CMD (CPY, DJJ, BWH), as well as this group. It aims to develop state of the art hardware and software for a European contender in the Stardent/Silicon Graphics market. The project is in several parts, and that with RAL involvement is to develop a board and associated software to enable quality 3D graphics interface to be developed. This part of the project is known as GRACE (GRaphics ACcelerator Engine), and RAL's role is to provide an Interaction Framework which will act as a vehicle for high-quality user interface design. The work covers a range of topics. These include:
The first phase of the requirements analysis was a broad brush approach at a wide range of possible uses the EYS could be put to, with a view to determining what range of facilities would be needed. The second phase was more detailed, but concentrated on far fewer possible applications. Everyone contributed to this part of the task, but with VAB taking the major part of the work. This phase of the project was one of the elements providing input into the development of a class hierarchy for an object-oriented toolkit, undertaken mainly by IY, with input from CPY. An assessment of the demands from a diverse set of sources graphics standards, other toolkits, environments such as Small talk, advanced graphics software such as Dore and RenderMan, as well as the requirements analysis, were used here. Implementation by IY on The Kit has begun on two class categories of Interaction and 3D. To enable the rapid prototyping of applications compliant with The Kit, an interface prototyper (Constructor Set) is being implemented within the OpenWindows environment using the NeWS toolkit (tNt). Support for lightweight processes and inter-client communication is being provided by integrating the necessary Chorus functionality into The Kit. A simple NeWS implementation of the Squeak user interface language was developed to provide a vehicle for experimenting with gesture.
BWH has looked after the building of the X server on the development environment provided (four months late!) by the Chorus Systems people. There has been a long learning curve associated with the Chorus distributed operating system, and how it relates not only to the X server but also to the specialised hardware. A deaf and dumb X server now runs on the Chorus machine - this can only receive events via an input file, and prints results via printf's rather than by displaying items on the screen. A sequence of more and more talking-and-listening servers is planned to enable the final system to be built on the target hardware/ software. DJJ has defined - a few times now - the interface the graphics hardware partners in Darmstadt should be working to, and has begun the simulation software development, with VAB's help, here at RAL. This will be built on top of that interface so that the higher levels of software can be built on Suns here at RAL before the porting exercise begins to the (indubitably flaky) hardware later in the (calendar) year.
MJP, with TC's help, has begun the evaluation of the DataGlove 3D input device, using the demonstration software supplied with the DataGlove. A number of ID staff have been put through the evaluation mill, with the objective, at this stage, of assessing the Glove against the mouse as an input device on a number of tasks. CAAG has interfaced the Glove to the Sun RS232 port, and DJJ has built on this work to show the orientation of the hand on the Sun screen.
Although names have been associated with specific parts of the project, it should be stressed that this has been an intensely cooperative project, with most people at least commenting at most stages. VAB has provided much of the day to day management on the project, with KR looking after the higher-level issues and the relations of the project externally, and to the RAL financial and administrative mechanisms.
Work on this project is reported fully in the UIG section of the report. TAW has been spending most of her time on this since April 1990, with various consultancy tasks being undertaken as necessary, mainly by KR, CAAG, and IW.
AVS designed and implemented a route display program for the Open Days. This provides a distributed program with six remote stations connected to one or more master terminals, with the software being run by an experienced user. A wide range of information is provided for the visitor, including the demonstrations to be seen at each location and some interest-to-demo matching. Frame-grabbed images of the site are displayed at relevant places. Data for the map used was taken from the Prime-based Medusa system and then adapted.
There has been substantial interest in the Intelligent Front End (IFE) by the energy modelling community. Numerous universities, including some in Europe and the USA, have asked for the software; both the Building Research Establishment and the DTI Energy Technology Research Unit are looking at it with a view to adapting it for their simulation work; and the standards committee of the Building Environments Performance Assessment Club, (a community-initiated(!) community club), is interested in it to provide standard interface to the wide range of energy models. Several companies have also expressed an interest, to the point of being willing to participate in follow-on research projects. Unfortunately, as it :is built on top of some proprietary software, there are problems with its distribution.
It is intended to a) carry out a proper field trial of the system, in collaboration with industrial partners, b) enhance the capabilities of the system, and c) explore further the 'conceptual mapping' aspects. To this end, proposals have been submitted to the CEC Joule programme and SERC, but, in spite of the very positive comments received, to date there has been no funding forthcoming. (In fact the Joule programme was cancelled for lack of funding; recently the Commission have asked us to resubmit the proposal.)
A follow-on proposal to EWS was submitted to the Commission, which built on the work done in the first part of the project. Unfortunately this was not funded, for reasons which seemed to be more political than technical.
KR has contributed to discussions on the preparation of this ESPRIT 2 proposal, reported fully under the UIG section. The proposal was agreed as technically suitable, but with only 23% of the requested funding. In the subsequent negotiations the RAL and Lancaster University bits fell (well, more were pushed, really) by the wayside.
As usual there are some items which do not fit cleanly into any of the above categories. They are presented in no particular order.
KR is Monitoring Officer for the PUMs project (now completed) and the UIDE project (just started).
Believe it or not, we are now close to signing a contract with Pixel Designs Ltd to market ten, an activity that began at least as long ago as August 1985.
The Alvey demonstration has been aired on several occasions including being video-taped for inclusion in the RAL video. SysIG staff are thus now stars of the TV screen, having conquered radio some time ago. AVS has all the demo software under her supervision.
AVS attends the User Services Meeting as one of the DD representatives, and reports from IDUUF to that meeting. KR attends the (infrequent) meetings of the Unix Liaison Meeting, of which CAAG is the secretary. VAB chairs IDUUF, and generally manages to keep the length of the meeting down to a reasonable length, no small achievement.
KR is on the organising committee for this conference, which aims to establish an annual conference on the X Window System in Europe.
Get everyone in SysIG/UIG running X11, and consequently all our tools on X11.
Staffing during the period September 1989 - June 1990
The above reflects the composition of the Group during the latter part of the period, following the return of DCS from sabbatical leave.
However, it is not far from the way the Group operated during the remainder of the period. By February 1990, it became increasingly obvious that more effort needed to be devoted to the ESPRIT-II project Elusive Office (ELO) and that this could only be achieved by TC temporarily relinquishing his post as Group Leader in order to work full-time on ELO. As a result DCS became acting Group Leader as from 1 March 1990, initially for 4 months but then on an ongoing basis.
The objectives identified for the Group at the beginning of the period were in the areas of UIMS, PHIGS and Human Factors as part of the EASE programme and participation in two ESPRIT-II projects EuroWorkStation (EWS) and Elusive Office (ELO). With the return of DCS from sabbatical leave, participation in a third project ARGOSI (Applications Related Graphics and OSI Standards Integration) was reactivated.
In the UIMS area, further evaluation of commercial UIMS was targeted with a particular emphasis on those UIMS which ran on top of X-windows. Referring back to the EASE Community on the first evaluation and identifying candidate sites for trials of selected UIMS were also included.
The PHIGS work was centred on the evaluation of commercially available PHIGS implementations together with consideration of the support tools necessary for successful use. Liaison with other ID and CCD Groups working on graphics should also take place.
The Human Factors work aimed at experimental work on the use of the DataGlove (in the context of EWS) and completing a study on the effectiveness of different types of menus.
The first stage of the UIMS evaluation, carried out on behalf of CFTAG, was concluded with a seminar held at UMIST in September 1989. The results of the evaluation were reported by TC, namely that none of PA Set, BLOX or Tiger could be recommended for use by the community or in the EASE Initial Environment. MJP presented a paper at the seminar and updated his RAL Report on the relative merits of selected UIMS.
LS did some further work on Tiger (including implementing a small interactive application for drawing project plan bar charts), as there were some promising features, but this work was terminated when, after a period with no response from Team Engineering Inc (the suppliers of Tiger), it was finally discovered that Team Engineering were no longer operating as a company. The fact that this period coincided with the San Francisco earthquake complicated communications and generated many theories as to their fate!
While CFTAG had been disappointed with the conclusion of the first stage of the UIMS Evaluation, in June 1989 they had recommended further work in the area with particular emphasis on UIMS ran on top of the X-11 windows system. MJP and LS produced a new candidate list of UIMS.
Open Dialog from Apollo Computer had previously been identified as a potential UIMS operating over X-11 but had not reached a sufficient stage of development for inclusion in the first stage evaluation. The evaluation of this product was assigned to Colin Walls at UMIST who was providing effort for this project during the remainder of the Prime support contract.
The list of UIMS to be considered at RAL was reduced in the first instance to USEIT from UNIRAS (important on account of the CHEST deal), Te1eUse from TeleLogic and SL-GMS from the Sherrill-Lubinski Corporation. These all run on top of X-windows. Initial evaluations of USEIT and SL-GMS have been completed by LS and the evaluation of Te1eUse is continuing.
GDH recently joined the group as a trainee scientist and during his six months project will further evaluate USEIT using the APPLE program provided by Bryan Colyer.
When these evaluations are complete, it is planned to report back to the community again.
Arising from his work on the UIMS evaluation, MJP was invited to participate in a panel discussion on UIMS at Eurographics 89 in Hamburg in September 1989 and further gave a half day tutorial presentation on UIMS techniques as the original presenter was unable to attend. MJP also was invited to give a talk at an IEE Colloquium on UIMS in November 1989.
PHIGS (Programmers' Hierarchical Interactive Graphics System) is an ISO standard for defining 3-D models and viewing them. This project is concerned with supporting PHIGS for the EASE community initially placing emphasis on evaluating different implementations. PK has been working on the development of an evaluation procedure.
PK initially had to teach himself about graphics and PHIGS (hampered by a lack of books and training material on PHIGS) before gaining experience with SUN PHIGS. All the SUN examples and tutorials have been made to work. PK has written programs to test the speed of 2D vectors, 3D vectors, 3D text and 3D transformations.
This work falls within the remit of the joint Computer Board/SERC Advisory Group On Computer Graphics (AGOCG) and DCS has begun liaison with this group. He submitted a paper with Ruth Kidd (DCSD) on EASE Support for Graphics Standards outlining possible future activities, ranging from assessment of implementations to support toolsets to providing guidelines on successful use.
Graphics is covered in several places within Informatics as well as in CCD and so there is a joint CCD/Informatics meeting to discuss common issues in computer graphics. PK has attended four such meetings during the period of this report.
Human Factors experimental work has been affected during the period by the HF Lab being used as an office for part of the time. Despite the cooperation of the occupant (YY), this has made the environment for experimental work less than ideal. This problem has now been resolved and it is planned to return the HF Lab to full and proper use.
For the ESPRIT-II EWS project, MJP has been evaluating the DataGlove as an input device from the HF point of view. He created test software and tested 15 subjects with the glove. The results have been analysed and a paper written. The EWS project is described in more detail in the SysIG chapter.
The menu style study described in the last annual report was completed and MJP has produced a paper which has been accepted for Interact '90 in Cambridge, UK in August.
MJP has also assessed the use of CD ROM technology using the Oxford English Dictionary on CD. On a different topic, MJP investigated the problem of excess noise levels caused by SUN workstations in offices and carried out initial studies on using sound absorbing materials to alleviate the problem.
It was satisfying that demonstrations based on the Human Factors work carried out by the group (and related work in SysIG) were selected for the RAL Open Days to be held in July 1990. MJP spent time tailoring the demonstrations and producing accompanying information material, with the result that some very professional demonstrations were produced.
With the return of DCS, the Group has been involved with three ESPRIT-II projects: EWS (EuroWorkstation), ELO (Elusive Office) and ARGOSI (Application Related Graphics and OSI Standards Integration). The first two started in January 1989 and the last in March 1989. EWS and ELO include staff from UIG and SysIG with the former project run from SysIG and the latter from UIG (TC is RAL Project Manager). ARGOSI includes staff from all four Divisions with DCS participating from UIG. EWS is described in the SysIG chapter.
ELO is concerned with the provision of the next generation of high power, portable workstations and was outlined in. the last annual report. RAL emphasis is upon a high-quality user interface and a framework for integration.
Work on the project was delayed owing to staffing problems for the first part of the period with the problem only being effectively resolved in March 1990 (see Staffing).
In spite of this the RAL contribution to the project seems to have been respected by other members of the consortium and the project has survived a number of review meetings. (There has been a presentation from RAL at each of these reviews.) During the year an extra (reviewer-suggested) task has been completed (namely, the ELOPE early prototyping exercise), good relations have been maintained with other partners (in particular, with CLS another software partner) and a major deliverable (A. 3 .1) completed. TC has been the RAL project manager and B Ritchie (SED), S K Robinson (SED), TAW (SysIG), KR (SysIG) and YY have all participated.
YY has contributed to the discussions on the functional split, the architecture and design strategy of the ELO integration framework. Since more than one person, each with an ELO system, may work together to carry out a task, issues related to computer-supported cooperative working (CSCW) are important. YY has been analysing MUSK (Multi-User Sketch Program) to study the structure and find possible places for modification and extension to gain expertise for this and other CSCW projects.
ARGOSI aims to advance the state of the art in the transfer of graphical information across wide area networks, using formal standards, and was outlined in the last annual report. R A Day (DCSD) is the RAL project manager.
DCS has participated in various parts of the ARGOSI project. He has worked on the classification of applications according to their graphics and networking requirements, analysing the results of a postal questionnaire distributed to over 2000 people working for organisations with applications using graphics or networking, and contributing to the discussions on constructing a classification framework. He has made contributions to the related Study of Services required. DCS has also contributed to the specification and design of the application demonstrator and to the design of the graphics services required for that demonstrator.
The course on Human Factors Aspects of User Interfaces described in the last annual report (as part of a project coordinated by ZGDV, Darmstadt) was finished off and the interactive software was completed and installed. The course was presented by KR (SysIG), MJP and TC in Darmstadt in September 1989. Material was prepared for a demonstration and handout on the COMETT course for the Eurographics 89 conference in Hamburg earlier in September. There are plans for a follow-on to the COMETT project and to repeat the course for the EASE community.
PK has worked with staff at DTI and RAL to encourage the use of appropriate standards in grants approved by the joint programme and to draft a proposal for a research project on Standards.
DCS spent the first half of the period completing his sabbatical year with the Division of Information Technology (DIT) of CSIRO in Australia. He worked at the Centre for Spatial Information Systems in Canberra. The visit was arranged as part of an exchange scheme with CSIRO DIT which had begun with the visit of Graham Freeman to Informatics Department in 1988.
DCS worked on the specification of KIPS (Kernel Image Processing Software). KIPS was designed to provide a simple, powerful, flexible and consistent environment for programmers developing image processing applications and for operators running such image processing applications.
KIPS consists of three defined interfaces and a dynamic set of tool kits and applications for operating on images.. The defined interfaces, each made up of specified functions together with their associated parameters and the data structures necessary to support them, comprise:
Tool kits, which can be specified by R&D or engineering groups using KIPS, are defined in terms of the interface functions and can be used for the subsequent building of applications.
A set of generic applications built using the interface functions will perform standard image access, handling, processing and display operations. Other applications to perform specific operations on images can be specified in terms of the interfaces and tool kits described above.
DCS was concerned with defining the overall architecture for KIPS, specifying the Kernel, its method of data access using a multi-dimensional format previously defined at DIT, and defining the display device model. In contrast to graphics systems, the central concept of KIPS was the image and operations on the image. Display devices were just handled as another kind of storage device with the special property that what was stored could also be viewed. More details are given in the paper on KIPS listed at the end of this chapter.
YY, in her role as Atlas Fellow, has spent a significant part of her time on project STRIP to identify factors that influence the usefulness of software tools. Currently some support tools for designers have failed to provide adequate support at the appropriate conceptual and operational levels. In order to provide adequate support tools, the environment in which a support tool can be used and the tasks, which a support tool can assist with, need to be identified. After study of current support tools and software practice, the application environment and task for STRIP has been identified as the design phase of software systems in an object-oriented environment.
One attraction of the object-oriented paradigm is to provide safe software reusability. One special development activity in designing an object-oriented software system is to understand existing classes of a chosen environment and to extend them. Thus, the initial tool produced by STRIP will help understanding of the behaviour of classes. The requirements analysis and initial interface design have been carried out. Architecture issues are under investigation.
TC has acted as Secretary of CFTAG and attended ECFE and CFTAG meetings.
TC and MJP have both been learning German during the period.
One proposal was put together for the second round of ESPRIT-II projects, entitled EuroCoOp. EuroCoOp was concerned with remote cooperative working supported by IT. The project proposal was given a good technical rating and funding was agreed. However, just before a Consortium meeting, the major UK partner (STC) decided that there was not enough effort for them and bumped RAL and the University of Lancaster from the project!
The work that was to have been done by RAL and the University of Lancaster has been put together as a new proposal, entitled Syncopate, and has been submitted to the Joint Research Council's Initiative in HCI/Cognitive Science.
Two proposals have been submitted to COMETT-II. One (entente) concerning training in IT and telecom standards is in association with CLS GmbH (Bonn) and the French Telecom University (Paris). The second is a follow-on to the project described earlier coordinated by ZGDV (Darmstadt).
The Unit runs the programme for the SERC/DTI Initiative in the Engineering Applications of Transputers including the London & SE Regional Support Centre.
Staff in post at end of June 1990:
1989-90: To carry out the Initiative programme in accordance with the Report of the CFC Working Party on Engineering Applications of Transputers, October 1986; taking into account the recommendations of a Market Survey, commissioned by the Initiative, and reporting during the first part of the year. In addition efforts will be made to target Continental Europe for the Awareness Programme.
1989-90: To establish the Centre as self-sufficient in funding with a good base of demonstration software, a regular supply of paying visitors and a full Course Schedule.
To put into effect activities based on the policy of the Initiative as defined by MRJ acting on the advice of the Steering Group, the Transputer Applications Management Group (TAMG). To carry through support activities already committed, and to monitor the effectiveness of the programme.
CB is Secretary to TAMG and RJF is secretary of the Directors Meeting, and was also Chairman of the Managers' Meeting. As a formal meeting this Managers' Meeting is now being replaced with specific interest meetings, usually to be held at Suppliers' premises. CB also liaised between the DTI and the Support Centres in drawing up a Project Monitoring scheme in line with requirements for DTI partial funding of the project. CB also arranged purchase of equipment for the Loan Pool and the Transputer Support Centres. Loan Pool applications have been dealt with by RJF and the resulting loans by LJR. The 'Public Face' of the Initiative is handled by MRJ at the official level, and TPM, principally through the Mailshot, but also by attendance at exhibitions, conferences etc.
CB has been primarily responsible for liaison with Suppliers of Transputer equipment, encouraging them to assist and benefit from the facilities offered by the Initiative. Typical activities are in negotiating sponsorship or special deals to have new equipment available in Centres at the lowest possible cost; support of the Initiative Training Courses through company's sales promotion etc and encouraging companies to submit technical material to the Mailshot for publication. Sponsorships negotiated in the period were:
The list price value of these is £255K. TPM also has strong involvement with companies in obtaining material for the Mailshot. Four Mailshots were sponsored by Companies during the period.
RJF monitors activities in the 6 Support Centres and fosters cooperation between them. RJF has been collecting Demonstration Software for use in all the Centres, and arranging exchange of training materials between Centres.
The Loan Pool coordination activity (RJF) is primarily the pre- and post-loan phase wherein loan applications are received and assessed against guidelines, and submitted to TAMG for approval or rejection, and also in the obtaining of deliverables in the form of reports for publication, and software for provision to the Software Exchange Library. The period has seen 36 Loan Applications of which 34 were approved.
TPM has been responsible for putting the Initiative in the Industrial and Academic Shop-Window. The Initiative Display Stand is to be seen at Exhibitions and Conferences around the UK and elsewhere. There are again two major events demanding much effort and support from the Laboratory and the Support Centres, (a) The PC User Show in June and (b) the Initiative Conference and Exhibition to be held at Southampton in July.
TPM continues as the Editor of the Initiative Mailshot, which has continued to grow in stature and acceptance by the Community. This year has seen the introduction of sponsored editions where the production costs are covered by a Supplier who gets direct promotional advantage as a result. The circulation has grown from 2100 to almost 2900 during the 9 month period of this report. It is interesting to note that 170 of these issues are sent to 116 Libraries to ensure even wider readership. The breakdown of the current circulation is as follows:
| Area | Academic | Non-Academic |
|---|---|---|
| UK | 1400 | 712 |
| Non-UK | 610 | 164 |
TPM also administers a video loan pool based on the Initiative's own video and one from Paisley College which is a tutorial on the Transputer. This has now made some 64 loans.
The Loan Pool was provided with an injection of new software and hardware products at the end of the Financial Year (value £250K), thus enabling us to offer academics the latest technology. This has resulted in an increase in activity albeit not quite on the scale expected. LJR has attended formal training in the use and maintenance of IBM PCs, which has helped greatly in the management of the Loan Pool, and was involved in some difficult software installations for Transputer boards in Sun Workstations.
MRJ is Head of the Centre and Chairman of the Operations Meeting.
The post of Manager was held by RP until September 1989, and by PJTS from then until March 1990. Since that date, the duties have been split between the main users and the Coordination Team, with CB putting in some 40% of effort.
LJR has assisted both with PC and Filestore management.
RJF is developing the available base of demo software and acts as Secretary to the Operations Meeting.
TPM maintains the Publicity programme for the Centre including a standing exhibit.
1989-90 has proven to be a rather difficult year for the Centre due mainly to a lack of continuity in its management. Despite this it has been possible to make modest earnings with some £12.8K coming from consultancy and training courses. The number of paying visitors has been very disappointing, as with the other Centres. In the RAL case we have had insufficient effort to address the direct needs of the Centre (eg marketing and involvement in long term consultancy work). However there is always a great interest shown by the many visitors (non-paying!) that we attract. The awareness component of the Initiative Programme as expressed through the Centre, can be seen to be very fruitful. The Centre remains a popular place for important visitors to the Laboratory.
The Initiative Exhibit has been shown at a number of events during the past 12 months, usually in support of a Transputer component at the Conference, and occasionally to market the Initiative per se.
SERC/DTI Initiative in Engineering Applications of Transputers, Joint UK/Soviet Workshop on Parallel Proce1ssing. MRJ - June 1990.
In 1990-91, to carry the programme forward, identifying how various activities might be dealt with in post-Initiative times ahead, and setting up any necessary liaison and training to enable this.
1990-91: To integrate the Centre with the main EASE activities of the Department whilst, at the same time, continuing to satisfy the original objectives set for the Centre (ie seek both paying visitors and consultancy contracts from, principally, local industry, and offer a full Course programme). The Centre should, in time, move to being a parallel processing Centre rather than just concentrating on Transputers to satisfy the broader needs of the EASE Programme.
The period of this report commenced with the appointment of a new Division Head, David Boyd. The group structure of the Division has remained unchanged. Both Chris Greenough and Chris Wadsworth were successful in achieving Grade 7 appointments to lead their respective groups. Conor Fitzsimons joined us from Trinity College Dublin to strengthen our effort in Computational Fluid Dynamics and Francis Yeung and Dayal Gunasekera (a Visitor from Swansea) left us having made valuable contributions to the work of the Division.
The Division has hosted visits from Dr Theo Kalamboukis from the University of Athens to work on the ACCORD project and from Dr Duncan Stevenson from CSIRO in Australia on an exchange programme, as well as offering a temporary home to two trainee scientists, Andrew Reay and Rob Woodward.
The principal strength of the Division remains the close interworking of physical modelling, numerical analysis, visualisation and parallel processing skills developed and applied through a number of engineering application areas. The current focus is on semiconductor device modelling, electromagnetics and fluid dynamics. Approximately two-thirds of the Division's resources have been committed to the EASE programme and the Transputer Initiative and the remainder to a number of projects under JFIT and ESPRIT.
Significant achievements within EASE and the Transputer Initiative have included:
Staff from the Division have also made significant contributions to the ESPRIT projects EVEREST, ACCORD, ARGOSI and EWS, and have secured funding for and started work on the BRIM and VAR projects under the JFIT programme. A further submission to ESPRIT during the period has subsequently received approval for the BECAUSE project.
All groups have now established series of technical reports, many of which will see a wider circulation as EASE Technical Reports.
The Division made the major contribution to a very successful visit by the SERC Chairman to Informatics Department in February, and contributed several demonstrations for the RAL Open Days in July.
The main purpose of the group is to promote the use of up to date mathematical techniques and software within the research community and to provide a focus of computational techniques and software from which the research community can request advice and assistance. This is achieved by being actively involved in research and development programmes, by providing software and expertise and by planning and participating in workshops and courses on mathematical software.
The major objectives of the group for this year have been:
During the year the Engineering Computing Facilities Executive (ECFE) was disbanded and the day-to-day monitoring of the EASE Programme taken over by the Informatics Department Heads Meeting (IDHM).
CG was a member of ECFE until its demise and has contributed to the discussion of the EASE Environment, the EASE Information Directory Service and the consideration of EMR contracts.
As a member of ECFE and Informatics Department, CG represented EASE on the IUSC Computer Algebra Steering Committee. He continues to provide a link between the IUSC Computer Algebra Support Project and EASE.
With the demise of ECFE two new working parties have been set up: the EASE Schedules Working Party and the EASE Environment Working Party. CG is a member of the former and JVA of the latter. The EASE Schedules Working Party has undertaken to review the EASE Schedules and produce an up to date document containing details of all the Computing Facilities Committee's recommended software, hardware and standards. The EASE Environment Working Party has undertaken to define and formulate a work programme that will lead to the initial EASE Environment with a number of applications demonstrating its usefulness.
Release 3 of the Library is now complete and with NAG Ltd awaiting distribution to the community. Documentation has proven the most time consuming task. Release 3 of the Library adds six new Level 1 Programs and twenty new Level 0 routines.
The new programs are:
These new programs add to the range of applications described in the Level 1 Documentation. They introduce the use of complex numbers, geometric nonlinearities (mesh movement) and nonlinear differential operators with solutions using point iteration and Newton-Raphson techniques.
The new Level 0 routines cover:
The use of these routines is illustrated in the new Level 1 programs.
To accompany the release, articles have been written for the Engineering Computing Newsletter and the NAG Newsletter.
A training course is being planned for later in the year.
Although Release 3 has taken much longer to complete than hoped developments for Release 4 of the Library have continued. The initial contents of this release have been drawn up and include four new Level 1 programs and their associated Level 0 routines. The programs will include:
Additions will also be made to the Level 0 Library. These will include:
In addition to the development of new programs and Library routines the documentation of the Library will be completely revised and typeset using LATEX.
The idea of gradient recovery in the finite element method is to compute, by post-processing from a standard piecewise polynomial finite element solution, an approximation to the gradient which is of a higher order of accuracy than that of the finite element gradient itself. This is of great importance to many engineering problems since the numerical solution is often computed in some potential whereas the quantity of real interest is the field associated with that potential. The gradient recovery process will allow engineers to obtain the best numerical representation of these physically important values.
The theoretical error analysis for two-dimensional problems has been extended to three dimensions and the software embodying these techniques has been designed and the first implementation produced.
During the year a new version of the Library has been developed for computers with concurrent architectures. The main goal of PARFEL is to enable users of the serial Library to transfer their programs to machines with a MIMD architecture with little modification.
PARFEL is an extension of the serial Library which includes routines that handle data partitioning and distribution over local memory MIMD systems such as the Intel Hypercube or NCUBE. The prototype version of PARFEL was written using a basic home-grown simulation of a MIMD machine on a Prime system which demonstrated the feasibility of the approach. With the demise of the Prime, work moved to the Sun and the use of the Intel Hypercube Simulator which simulates the behaviour of both the Mark I and Mark II Hypercubes. This transfer has lead to improvements in PARFEL and work is in hand to implement the Library on the Intel Hypercube Mark II at Daresbury which will allow benchmarking to make place.
As Transputer based systems are very important to the engineering community a Transputer version of PARFEL is being developed using 3L Parallel Fortran. This version should be ready early next year.
EVEREST is a four-year project supported by ESPRIT which is investigating suitable computational algorithms for the analysis of semiconductor devices in three dimensions, and developing a software suite implementing the most effective of those algorithms.
The partners in this project are drawn from some of the major industrial and academic research and development teams in Europe. They are:
The activities of the project have been divided into five Workpackages: Physical Models and Validation (SGS), Discrete Problem Formulation (Philips), Mesh Generation and Refinement (UCS), Solution Procedures (Philips) and Software Suite (RAL). The coordinators of each Workpackage are shown after each title.
The major objectives of the project during this year have been:
The main focus of RAL effort has been in the implementation of the three-dimensional adaptive solver. CG has continued to be the coordinator of the software development within the project and RFF has played an important role in implementing the EVEREST Suite on the Stellar GS 2000.
RFF and DG have continued developing the Solver Module of the suite, DG having responsibility for the discretisation and linear algebra, and RFF concentrating on the physical models library and integration.
RFF has implemented the EVEREST Suite on the Stellar GS2000 making use of the multi-processor nature of the machine. This has required some modifications to the programs but the performance gains have been good. However much work is still required to make effective use of the four processors in the linear algebra routines.
In late November the EVEREST Project made a number of presentations at the ESPRIT Conference and had a stand in the Conference Exhibition. Both CG and RFF were involved in the preparation of the conference papers and the exhibition display posters and software.
The main focus of the presentations and the exhibition were the advances in numerical simulation made by the Project. The University of Bologna made an excellent presentation on their work in Workpackages 1, 2 and 4 as implemented in the simulator HIFIELDS-3D. As director of Workpackage 5 CG made a presentation on the developments of the EVEREST Suite.
In the Conference Exhibition the display stand was centred around a live demonstration of the EVEREST Suite on an FPS 350 Graphics Super-Workstation (really a Stellar GS 2000). The EVEREST Suite was shown running a selection of problems solved during its development in a continuous cycle. The demonstration was supported by display boards and many handouts.
Although the number of visitors to the stand was small (as one might expect for such a specialised application) the display and demonstration was thought very successful by the Commission.
THRUST is a training course on semiconductor simulation software supported by the European Community through the COMETT Programme. As with all acronyms THRUST is well and truly contrived: Training in Hands-on experience in the Real Use of Semiconductor software Technology.
The purpose of the course was to introduce the participants to the capabilities and limitations of state-of-the-art process and device modelling programs. The emphasis was on the practical aspects of using simulation software and there was ample opportunity for the attendees to evaluate commercially available software. The course was made up of a number of formal lectures presenting important aspects of process and device simulation and a large number of practical sessions in which the software was used to solve a selection of pre-prepared problems. These problems were designed to illustrate the important advantages and limitations of the simulation packages.
The course was organised by the Professor Sean McKee of Strathclyde University with help from Dr A.J. Walton of Edinburgh University. CG was a member of the founding committee and was involved in developing the course programme.
RFF attended this full course as a member and CG attended the final day to give a presentation on the EVEREST Software Suite. During the week RFF mounted the suite on the computing system being used by the course and gave demonstrations. While the course was primarily aimed at two-dimensional simulations, the EVEREST Suite was the only three-dimensional simulator available during the course.
The course was intended for process and device engineers working in industry and about 15 people attended.
During this year the ESPRIT Project ACCORD comes to an end. The Department has been involved in two parts of this project: the Vector Field's demonstrator and the APPEAL Library. Members of the Parallel Processing Group have been involved in the implementation of the Vector Field's demonstrator ABEL. This is a parallel implementation of a boundary element package on the IBM 6150 with Transputers attached. This demonstrator was display at the ESPRIT Conference in November 1989.
CG has continued to be involved in the APPEAL Library of ACCORD at a low level. APPEAL is a library of numerical routines addressing the solution of large sparse systems on vector and concurrent machines. CG has been active in developing standards for this project and in testing each new version of the serial library on the Cray and Stellar. The major effort of the project has been toward vector machines, in particular the FPS M64 series although the base version of the library provides some very useful numerical routines for serial architectures.
The APPEAL library will be made available to the SERC engineering community during the next few months.
EVEREST and TAPDANCE are two major software suites resulting from ESPRIT and ALVEY projects. Under the EASE+ initiative these suites are being supported for the microelectronics community.
In a response to a recent article in ECN over thirty academics have requested access to this software. Plans are in hand to make the software available and to provide suitable training courses in their use.
As a result of the ESPRIT project EVEREST five members are attempting to set up a User Club to fund the support and development of the EVEREST Suite. This activity would be based at RAL and would complement the effort being funded by EASE+.
Enquiries to simulate specific device types have been received from a number of industrialists and academics and these applications are providing very useful tests for the programs.
The two-dimensional mesh generator AQUABAT has been embedded in TAPDANCE to complement the existing ESCAPADE generator. This has allowed the investigation of mesh adaption using TAPDANCE and the device simulator DEVMOD.
Over the next year the Mathematical Software Group will continue its programme in research and development and the provision of software and expertise to the EASE community.
With the completion of Release 3.0 of the NAG/SERC Finite Element Library and the two ESPRIT Project EVEREST and ACCORD the Group will promote the results of these projects as well as continuing new development within the ESPRIT Project BECAUSE.
The EASE Programme will play an important role in the development of the Group's programme, particularly with the definition of both the EASE Schedules and the EASE Environment.
Specific objectives for the year will include:
Most of these objectives will produce both software and reports which the Group will make available to the engineering community.
The Engineering Applications Group is concerned with the assessment and exploitation of new computer architectures for engineering design and analysis. This may include supercomputers such as the CRAY X/MP; the STARDENT GS2000 and TITAN workstations; and parallel machines like the AMT DAP and transputer arrays. The experience gained in this work is used to provide guidance to members of the UK academic community within the EASE framework. Other projects are in the form of consultancy work and make use of the Group's expertise to provide assistance outside the SERC Engineering Board area.
Staff in post at end of June 1990:
The main EASE projects during the past year have been:
This work comprised taking the 3D electrostatic/magnetostatic package TOSCA and implementing it on the STARDENT computers, whilst making use of the vectorisation and high speed graphics rendering capability as much as possible. A version of TOSCA has been mounted using X-Windows. However, further work is required to achieve faster graphics using the AVS visua1isation package.
A study of solution of sets of linear equations involving dense matrices has been carried out and applied to the BIM2D magnetostatic program. A new fast matrix multiply algorithm has been developed using a recursive procedure.
An evaluation of the DAPIX visua1isation graphics has been carried out.
The work on the DAP has now been brought to a conclusion and the machine is being transferred to Particle Physics Department.
The Group have participated in the TEAM (Testing Electromagnetic Analysis Methods) workshops. This is an international collaboration to prepare and run a series of benchmark problems for verification of electromagnetic computer programs. A Workshop and Seminar was organised by the Group and was held in Oxford in April.
Further development of the RFQTRAK space charge program for RFQ accelerators has been done under contract from Los Alamos. This uses the CRAY X/MP computer. It is run in conjunction with an interactive graphics pre and post processor written by Vector Fields Ltd.
This was a study of the magnetic field distribution in an octupole magnet designed for insertion into the ISIS ring. Both PE2D and TOSCA packages were used in this analysis.
This was a pilot study to assess the suitability of the DAP for this package which simulates the interaction between a time varying wave in space and a collisionless plasma containing 100,000 electrons which undergo Brownian Motion in phase space. First indications were that the DAP is well suited to this type of problem.
To continue to assess the best ways to utilise the latest parallel and vector machines in the mounting of engineering applications packages. Emphasis will be on finding the best numerical techniques for each system. The hardware involved will include a wide range of architectures, for instance the STARDENT, FPS workstations and transputer arrays.
The Parallel Processing Group (PPG) undertakes projects and development work on the systems aspects and programming techniques of parallel processing, as well as monitoring and assessing developments in the field generally. Advice and assistance is also given on porting applications to parallel systems. PPG also supports the programme of the SERC/DTI Initiative on the Engineering Applications of Transputers and of its London and South East Regional Support Centre located at RAL, both of which are coordinated in separate Units in ID.
HKFY moved to Daresbury Laboratory on promotion to Grade 7. CPW was promoted to Grade 7 in February 1990.
Current interest is focussed mainly on scalably parallel architectures such as ones with distributed memory, eg the transputer and hypercube machines. A medium-term goal is the integration of the transputer and/or other parallel systems into the EASE programme. The "ease" with which this can be achieved may well depend on the rate at which the community develops the ability to exploit parallel processing without the need for explicitly parallel programming. The long-term goal is to provide an effective environment for developing applications on and for parallel systems.
Part of the Group's work centres on the porting of applications to parallel systems, with particular aims both to gain practical experience of the most useful parallelisation techniques and to consolidate an understanding of user needs for an effective applications environment for parallel systems. This work is typically undertaken jointly with an applications team, eg with other groups internally, or via Transputer Centre contacts, or on consultancy to an external applications party.
The major objectives for 1989/90 have been as follows (in no particular order):
Since September 1989 PPG has been providing additional effort for the ESPRIT EWS project run by Ken Robinson in Design Division (see that Division's report for a fuller report). This has principally involved BWH and DJJ, each part-time, with a small amount of input from CPW (on the class hierarchy for the project).
BWH has constructed an X11 (version 4) server for SUNs, using the MIT release, and evaluated performance on a SUN 3/80. The CHORUS operating systems, providing a distributed real-time kernel with support for multiple threads and IPC, was installed on SUNs and test programs written to try out the message passing and threads mechanisms.
BWH then ported several Xclients and Xlib onto CHORUS, enabling the graphics requirements of clients to be fulfilled by a remote X11 server on a SUN. The CHORUS system was shown to be reasonably compatible with Unix System V.
The standard X11 server was then ported onto CHORUS, which required production of a special access frame and simulation of input devices and their output. Xclients were created to test the operation of this server.
DJJ first surveyed the major graphics standards to determine those most relevant to the (3D) needs of EWS, then designed a hierarchical graphics interface to support the chosen standards in a consistent manner. DJJ has organised RAL's discussions with FhD-AGD (Darmstadt) to determine the interface between respective parts of the project and identify the functionality of the 3DGRP hardware and software. The latter has been tested locally on a SUN by writing a simulator for 3DGRP.
Last year's report deferred estimates of when this project might eventually receive its formal offer/announcement from IED. This eventually happened in December 1989 and project formally started on 1 April 1990.
The project was given the short name BRIM, signifying its four partners: Brunel University, RAL, Intercept Systems and Meiko. The aim is to develop the techniques needed to detect and extract parallelism from Fortran programs for multi-transputer target system.
Progress to date has focussed on the requirements for an intermediate language PDF (Parallel Distribution Format) for the project. The project at RAL involves CPW and DJJ, with Stuart Robinson (Systems Engineering Division) as a research consultant.
This Joint Study involves the IBM 6150 workstation and the transputer. Three 6150s have been provided by IBM on loan for the duration of the project (November 1988-0ctober 1990). The general aim is to explore the potential for enhancing the capabilities of an advanced graphics workstation (6150) with a reconfigurable parallel compute server (transputer). Work during 1989/90 has concentrated on porting and parallelising further applications after successfully completing the first example BIM2D in 1988/89.
Two substantial packages have been ported: Apple, and Onda. Both are moderate-to-1arge sized packages (20,000-50,000 lines) which previously ran on minicomputers and/or mainframes. APPLE is an electromagnetics design package similar to BIM2D. Onda is a river modelling package developed by Sir William Ha1crow and Sons for area water authorities.
BWH helped Vector Fields and Brian Colyer to generate an Apple demo for the ACCORD project at the ESPRIT Conference in Brussels in November 1989. The Apple port follows a similar approach to that for BIM2D, with the I/O and graphics remaining on the workstation and the transputers (32 currently) carrying out the compute intensive tasks such as (parallel) Gaussian elimination and generation of contour maps. The demo has been shown subsequently to many visitors, including the SERC Chairman, members of the RAL Visiting Panel, and the Industrial Directors of the three ERCIM organisations.
BWH has continued to refine the demonstration version. The graphics have been enhanced from using the ww interface to a standard XII interface, hence XApple as name of current demo. A four transputer version has also been generated for occasions when the larger 32 transputer system is unavailable.
HKFY undertook a major port of the Onda package, with some assistance from Halcrow in the form of updates to the application-specific routines (which are used unchanged). A parallel strategy was devised, then implemented and tested. Several problems were encountered particularly in the limited amount of memory per processor on the system used for testing. With some re-thinking, a working version was completed for a four transputer system with 1 MByte memory per processor. The strategy followed and some performance figures are reported in a paper presented at the Transputer Initiative Conference (TA90) in Southampton in July 1990.
With the move of HKFY to Daresbury, PPG has little effort remaining for further work under the Joint Study. Fortunately all the goals of the Joint Study have been achieved with six months still to run. A final report summarising the project remains to be completed by October 1990.
Further work with the 6150 has in any event become redundant. Towards the end of the period reported IBM announced a new RS/6000 range of workstations with about ten times the performance of the 6l50. One has been ordered and delivery is now expected in August 1990. Further work using a workstation/transputers combination will continue with the RS/6000.
Two lessons have become increasingly apparent from the work with APPLE and Onda. First, the speed of the interface between the 6150 and (the first) transputer (50 KBytes/sec) is an increasingly limiting factor both on overall performance and on full effectiveness of the demos generated. Fortunately higher performance host interfaces are now available for (some) transputer boards. One is a must for the RS/6000. (Indeed, higher performance transputers would not go amiss either, by at least a factor of 10!). The second lesson is that substantial applications require lots of memory when ported to transputer systems (particularly in the absence of virtual memory on transputers). A minimum of 4 MBytes per processor is becoming the norm in this community generally.
Logically this may be divided into support for the Transputer Centre and support/advice to the Transputer Initiative Co-ordination Team, though in practice it is sometimes difficult to distinguish the two roles.
PPG supports the Centre through provision of courses, participation in discussions with prospective industrial customers, consultancy for customers, and day-to-day support to users including dealing with many telephone, email, and postal queries.
Courses have continued at approximately the same level built up to in 1988/89. The two existing course Transputer Awareness (half day) and Exploiting the Transputer (one day) have each been given four times. DJJ has developed a new course Parallel Fortran (2 days) which has now been given three times, with BWH helping in the practical sessions. In addition a 3-day unit for the Atlas trainee scientists was given, which resulted in a further trainee (RJW) joining PPG for a 6-month project.
Among the companies with whom discussions have taken place are:
The asterisks indicate those companies with whom there has been specific follow-up this year. A number of academics have also visited to discuss their applications:
User support, and answering queries, has at times threatened to swamp the effort available this year, given that those providing it (DJJ, BWH) have increased commitments to other projects. Measures have needed to be taken to ensure that the time taken up is kept to a reasonable proportion.
All members of PPG provide support and technical advice to the Co-ordinator on miscellaneous matters from time to time. In particular this year: DJJ has opened discussions with Birkbeck College about the pull-through of their work into demonstrators and is liaising with Raymond Fawcett on this and other demonstrators for the Initiative; CPW provided technical liaison for completing the QUB/Brunel/Liverpoo1 EMR contract; CPW has continued as a member of TAMG; BWH has taken over (from HKFY) the support for Apollo transputer kit in the Academic Loan Pool; CPW gave presentations about the Initiative to the NAG Users Association annual meeting in York, and to the IEE SERTS Conference in Cirencester.
A Workshop on Software Standards for MIMD Machines was organised for the Transputer Initiative at Cosener's House on 1-2 March 1990. As with the previous Workshops, this took the form of invited presentations followed by three parallel subgroups (Language, Libraries, Systems) and reports back to a general session. CPW was programme organiser and chaired the Languages subgroup.
CPW compiled most of the workshop Report with inputs from HKFY, BWH, Cyril Balderson, and Raymond Fawcett who acted as notetakers in the subgroups. The major recommendation (being implemented) is the establishment of two Working Groups on high-level and low-level standards respectively.
The range of demonstrations described in last year's report continue to be available and most are demonstrated regularly. Two new demonstrations have been added this year:
The first was developed by BWH with Vector Fields (see 4.4.5). DJJ developed the second at short notice - assisted by midnight oil! - for an exhibit at Leicester Museum of Science and Technology where it ran unsupported (by us) for a month. This demo cycles through a number of features embedded in DDJ's previous interactive demo. The demo even boots itself automatically when the host PC is switched on!
Among those to whom demos have been given, in addition to courses and the companies mentioned above, are:
PPG's involvement in the EWS project has implied rather less effort being available for specific transputer developments, though a number of tasks have been undertaken. Two Atlas trainees joined PPG for 6-month projects (one continuing to the end of 1990).
XCon is a Norwegian company based in Oslo who have developed a transputer-based image compression algorithm for a client company, Pablo AS, in conjunction with a proposed geographical data base for the city of Oslo. The XCon prototype uses a custom-built transputer board, incorporating FIFO-buffers in the host interface for higher bandwidth, and was hosted by a PC running Xenix.
Following discussions with the Centre Manager, CPW, and BWH, the Transputer Centre were contracted by XCon to port the prototype to an Apollo workstation as host (with a view to using the richer user interface capabilities in further developments). BWH undertook the work, porting the host server/driver and incorporating the driver into the Apollo's particular device driver system. This was completed in 15 man-days exactly as estimated to XCon!
DJJ collaborated with Patrick Pope of T2 Systems, a user of the Transputer Centre, in converting disk driver software to C to run under the GENESYS system on Transtech's NT1000 boards (ex-Niche) in SUNs.
DJJ has polished his prototype parallel worm program for exploring transputer networks. A paper is being prepared for submission to Transputing 1991, the first world conference of Transputer and Occam User Groups to be held in the US in April 1991.
AMR started a 6-month project in January 1990 extending BWH's idle-time monitoring tool, in particular enhancing the user interface. At the time of writing the final report of the project is being completed.
RJW joined PPG towards the end of the period reported and will be working on configuration aids for transputer programs, supervised by DJJ.
PPG received two products, the 3L parallel languages and Helios-PC, for evaluation under this programme operated by the University of Southampton for the DTI. An interim report mainly addressing our experiences with 3L Parallel Fortran was written for a meeting of the programme participants in April 1990. Shortage of available staff effort has limited evaluation of Helios, though CPW has written a book review for a journal.
Achieving a stable environment for portable parallel software, across a significant range of parallel architectures, is increasingly seen as critical for the widespread exploitation of parallelism. CPW has continued to spend much time cogitating these matters, though rather less this year in getting thoughts down on paper. A position paper was provided for the Workshop in March (see 4.4.6.3) and talks given at the IEE SERTS Conference in Cirencester and at EASE 90. The Working Groups being established following the Workshop will be addressing many areas where standards can help.
PPG joined with Visualisation Group in the IEATP collaborative PRIPS proposal Parallel Reconfigurable Image Processing Systems submitted in December 1989 led by NAG. The proposal was technically approved quite rapidly but then ran into difficulties when one of the industrial partners withdrew. Attempts to rescue the proposal are continuing.
CPW and HKFY each attended meetings in the lead-up to finalising the proposed ESPRIT programme on the European Microprocessor Initiative (EMI). A joint position paper was presented by HKFY at one of the EMI Workshops in Brussels in February. The programme has apparently now been approved in outline, though no further details are to hand yet except that the name has been changed to the Open Microprocessor Initiative (OMI). CPW has been invited to join an Industrial Working Group tasked with defining the Work Programme for a Call for Proposals.
Some details of the long-awaited DTI/SERC Parallel Applications Programme (PAP) emerged towards the end of the period. CPW assisted DRSB and MRJ in the preparation of RAL's part in Oxford's bid. Response to the outline bids is awaited.
Further discussions on common interests and possible joint activities with the IBM UK Scientific Centre, Winchester, are being pursued.
The general directions outlined in 4.4.2 will continue with increased emphasis on the techniques and methods, including standards, needed to bring about a mature parallel programming environment, and the integration of parallel processing systems into EASE.
Specific objectives for 1990/91 are as follows:
The group was greatly strengthened by the arrival on 17 April 1990 of Dr Conor Fitzsimons (CJF) to join Manjit Boparai (MKB) and Alan Bryden (ADB), who is the group leader.
CFD is a major applications area supported by several committees of Engineering Board, viz Electromechanical Engineering (EMEC), Process Engineering, and Environment. There is also funding from-Science Board for some fundamental developments.
The overall objective is to provide the computing infrastructure required by the Engineering Board CFD community to enable them to exploit modern computing techniques, and to keep the community informed on what is available. The mechanism to achieve this is to establish a Community Club in CFD which will help to set priorities, identify particular topics for investigation, and disseminate the results.
The CFD programme under EASE was originally proposed jointly with EMEC who agreed to fund a CFD Co-ordinator. However there were delays in appointment of the Co-ordinator because of the proposal within EMEC for a LINK Programme in CFD. It was eventually decided not to wait for the EMEC Coordinator to be appointed but to proceed with setting up the Community Club which held its first meeting at RAL on 22 March 1990. It was attended by more than 60 people.
A Steering Group for the Community Club has been set up with representatives from the three Engineering Board Committees funding work in CFD and from Science Board. There are also members to provide links to ERCOFTAC (the European Research Community On Flow Turbulence And Combustion), and MOD. It will hold its first meeting on 17 July.
There are now more than 130 people registered with the Community Club.
There has been close contact with the Visualisation Group, which has led to two contracts being awarded by CFC on Visualisation in CFD.
One is a pilot development in use of advanced visualisation systems in real engineering applications in CFD. An invitation to bid was issued in January 1990, and eleven bids were received. The contract was awarded to Dr Weatherill of the Civil Engineering Department, University College, Swansea. It will run for two years from 1 October 1990.
The second concerns the placement of the Ardent TITAN computer. Following an invitation to bid in August 1989, it was agreed that the TITAN would be placed in the ABACUS Unit at Strathclyde University for six months, and thereafter in Professor Gosman's group in the Mechanical Engineering Department of Imperial College London.
On the technical side MKB extended some of the work carried out in her MSc thesis by writing an interface to enable FAMResult (the post-processing section of the FEGS software suite) to be used to visualise output from a flow in a cubical cavity.
It was agreed at the Community Club meeting that a series of workshops will be held. The first is on 19 July at RAL on Visua1isation in CFD, when many of the bidders for the contract mentioned above will describe their work. Seventy-five people have replied to attend the workshop.
MKB is investigating the use of the PHOENICS code in a variety of problems in CFD. The aim is to determine the feasibility of implementing algorithmic changes in a commercial code like PHOENICS without access to the source code. It is expected that this will lead to a set of criteria against which other commercial codes can be evaluated.
During the period MKB has investigated the effect of changing the following items: Boundary Conditions, Physical Properties (density, laminar kinematic viscosity, turbulent kinematic viscosity, temperature), Dependent Variables, Array Dimensions, Discretisation, and Grid Generation.
In addition, ADB took part in EASE visits to Queen Mary Westfield College, London, Polytechnic of East London, and Thames Polytechnic, as well as several visits to other HEIs to discuss their CRF requirements.
Internal CFD Group Notes:
MSc Thesis:
This work is already under way. It will be completed by the end of the calendar year.
The Advisory Group recommended that encouragement should be given to the use and development of common software in CFD. The Abingdon Workshop on CFD recommended that there should be the provision of a variety of mesh generators and a greater pulling together of effort in complex postprocessing.
Dr Yuan Jiansheng (Beijing) is visiting RAL for one year to write a geometrically flexible three-dimensional mesh generator. It is planned to collaborate with Dr Yuan in this task, and thus ensure that the resulting software will be of use to the CFD community.
How we intend to address the recommendation on post-processing is discussed under the heading of Visualisation.
Paragraphs (a-c) below outline the programme of meetings for this year which have been agreed with the Steering Group. Paragraph (d) mentions those which have been agreed with the Steering Group for next year. Paragraph (e) concerns the collation of results from the questionnaire circulated to Community Club members earlier this year.
The main activity is to monitor the work on the EMR contract in Visualisation in CFD at University College, Swansea and to monitor the placement of the Ardent TITAN at Imperial College, London. Part of this will involve ensuring that the overlap between these two contracts is minimised. It is intended to arrange a meeting with UCS and Imperial and the Visualisation Group at RAL in September to outline complementary work programmes.
RAL staff will work closely with UCS to ensure that the results of the EMR satisfy the recommendation from the CFC that the visualisation system be capable of interfacing to a wide range of CFD systems.
The staff in Visualisation Group are:
Duncan Stevenson from the Division of Information Technology, CSIRO is with the group for 6 months from June 1990.
The group was formed with a brief that was not previously explicit in the department - to be responsible for the visua1isation of complex data and models. The main objectives therefore were to:
Increasingly the group will be investigating flexible visualisation systems, producing software where needed and cooperating in solving specific application problems. Until now, the group has been solving specific problems with lower level graphics systems and windowing system libraries (see "Visualisation Support for CMD projects" below).
AVS2 was installed on the Stardent ST2000 (ex-Stellar) (JH) and has been made to work with the spaceball (JH). Some experience is being gained with it. It has the potential to be a good flexible system for visua1isation and Stardent have concluded deals with other suppliers including DEC, which promises well. Since then additional modules from Stardent have been installed.
However, since the field is still volatile, information on other visualisation systems is surveyed and disseminated. JRG has given presentations at CG89 and at the SERC workshop at Daresbury in June 1990. JRG, JH and RP gave a tutorial at EASE 90 (see below).
JH has continued to gather information on software, claimed to be visualisation software, from both commercial and public domain with a view to future use at RAL or by EASE or other projects.
The Stardent ST2000 provides XFDI, PHIGS+ and AVS graphics and visualization packages. So, when the EVEREST package was ported to Stardent ST2000, its postprocessor module which uses RAL-GKS graphics library could not work. RP was asked to solve this problem. There were three alternatives:
The decision was made to leave EVEREST postprocessor unchanged and to concentrate on GKS implementation. Since the RAL-GKS X-driver was not ready and there was not sufficient time for porting, RP opted to simulate GKS calls using PHIGS+ routines. 44 GKS routines, used in EVEREST postprocessor, were implemented through PHIGS+ calls and successfully tested. This enabled completion of EVEREST's porting to Stardent ST2000.
When the TOSCA post-processor was ported to the Stardent ST2000, JH provided the graphics support. TOSCA on other systems uses the GINO graphics library. The idea was to change as little TOSCA code as possible. Initially, emulating GINO using PHIGS+ was attempted. This showed up several deficiencies in the implementation of PHIGS+ and was not continued. The final (and successful) emulation method was to port Brian Colyer's package which emulates GINO using ww and to port the X11 version of ww. X11 is not precise in its requirements for colour support in an X server and indeed problems were encountered and solved.
JRG, JH and RP prepared and presented the EASE 90 tutorial on Data Visualisation. The tutorial was split between them. JRG gave the Introduction to Visualisation in Science and Engineering. RP presented Hardware and Basic Graphics Software for ViSC. JH presented Visualisation Software. 25-30 people attended the tutorial.
The tutorial included slides and videotapes. The single video for RP's presentation was a compilation from a variety of sources in both PAL and NTSC.
The tutorial also included demonstrations of AVS on the RAL Stardent ST2000 which was at EASE 90 for the exhibition.
The transfer of the ST2000 was organised by the group.
The Alvey project has now been completed and written up as a final report. However the work to get the library into shape for the release of the mark 1 version has continued. A subset of the C code has been out on test release to a small number of sites. This has produced useful comments and action is proceeding to modify the code in order to satisfy the testers. The libraries have been written up and it is intended that this will be issued as a joint NAG/RAL report. Even this process led to useful feedback and structure definitions were simplified. Work is proceeding in two areas. One is to provide wrappers to allow the Fortran library subroutines to be called from C. The other area is an investigation of possible user interfaces to the library.
In December a presentation was made to the visiting IT research review committee (JRG).
It was decided to devote part of the IPAL Advisers meeting on 23rd March to the subject of standardization of image processing. Dr Krolak of the University of Lowell, who is driving the American effort was in Europe and was willing to speak. The advisers meeting was subsequently cancelled but the opportunity to discuss such a key issue seemed too good to miss. The majority of the original attendees were interested in a meeting devoted to standards issues and after being given wider publicity, the meeting went ahead chaired by Chris Cartledge. A useful and at times quite lively meeting followed and it was clear that there is a strong UK interest in this matter. One consequence was the setting up of an BSI Imaging Panel under IST/31/-/6. This panel is chaired by Adrian Clark of Essex University and held its first meeting on 10th April 1990 at the BSI. This meeting discussed the proposed new work item N427 and made recommendations as to what the UK response should be. (RM attended) The meeting looked for possible UK input and suggested that the report on the NAG/SERC Image Processing Library could be used for this purpose. The report will be tabled at the Imaging Study Group meeting in Ottawa.
A proposal was submitted to the SERC/DTI JFIT programme, the other partners being NAG, Meiko, University of Essex and MRC Edinburgh. The proposal was approved by the appropriate JFIT committees, but Meiko have withdrawn from the consortium. Other solutions are at present being sought but the rules regarding the funding of industrial resources committed to the project make this a frustrating exercise.
The proposed work aims to exploit dynamically reconfigurable parallel processing in the context of image processing.
This is a collaborative project involving BAe, RSRE, University of York and University of Surrey. Progress through the various levels of SERC-DTI approval has been painfully slow. This was mostly due to the ambiguous position of RSRE in its transition to agency status.
In April a SERC-DTI grant to the value of £158,000 was finally awarded to RAL with ERR as Principal Investigator. Work has now begun on the project with some studies of relaxationa1 dynamics on Hopfie1d networks. A paper describing this work is in preparation.
Several items were completed in connection with this project although it officially ceased on 30 March 1989.
In December ERH made a presentation to the visiting IT research review committee. In addition, papers describing two pieces of research work conducted as part of the project were also completed.
One of these described some work by ERH which demonstrated the Bayesian basis of a hitherto heuristic but highly popular vision algorithm - Canny's edge detector. This work is important for several reasons besides the novel interpretational basis that it provides. Firstly it endows Canny's algorithm with an adaptive capability. Secondly, and perhaps of more importance, its idea of applying the hysteresis algorithm iteratively. This has been shown to have some very tangible performance gains.
Finally, a paper describing work on discrete relaxation was selected as a long paper at the IEEE Computer Society 10th ICPR in Atlantic City, USA. This represents a real achievement since the paper was one of the top 20 of the 700 conference contributions.
The project is aimed at developing a photon counting detector supporting detector development in the SOHO CDS experiment in particular, and also work in a range of other application areas. The main thrust of the work over the last year has been directed towards developing a prototype photon counting detector demonstrating the performance achievable with this device and paving the way for a future flight system. The prototype detector has been built in the laboratory and the Transputer processing system and software developed allowing on-line data analysis and real-time image processing.
Exciting preliminary results showing the potential of this detector have been published. It has been demonstrated that the resolution is limited only by the channel spacing in the curved microchannel plate used in the image intensifier. The prototype system provides a valuable tool for the study of the image intensifier and the enhancement of the data processing algorithms. A collaboration with Kings College London has been instigated to study some of the characteristics of microchannel plates.
Future work will be aimed at continuing to improve the performance of the system through both software and hardware development and to produce a device for field testing on a telescope as well as extending the application area into UV and X-ray detection through development of the intensifier and CCD system.
In connection with the photon counting project, but supporting more general detector development, we have also been involved in setting up a general image processing facility, based around Transputers, for real-time detector applications.
The last few years have seen the emergence of significantly more powerful workstations available to scientists and engineers. Superworkstations, the term we have selected from the many hype terms on offer, actually expresses the situation quite well, as the computational and the graphical aspects of a workstation are uprated. Typically the cpu includes vector processing capability, although some systems are competitive even without this. The graphics processor typically allows 24 bit graphics and an accelerator for 3D.
The work started in November 1989. The purpose was to report to CFTAG in June 1990. After the project started, we were requested in addition to provide an interim report to CFTAG in April.
The strategy, used in other hardware assessments, of using a questionnaire and benchmarks was adopted here also. To help establish the criteria for inclusion in the assessment, RP compiled a table of manufacturer's supplied product specifications.
RP designed the questionnaire and analysed the results. The questionnaire for Superworkstation Assessment was based on the one designed for the Workstation Assessments. We put more emphasis on the graphics part of the system and took into account new features offered by superworkstations. The questionnaire was sent to five suppliers and altogether six different systems (Stardent's ST2000 and ST3000, Hewlett- Packard's DN10000, Silicon Graphics' 4D Power Series, Megatek's Sigma 70 and Evans & Sutherland's ESV series) with all their variations were covered.
RP obtained and ran a set of benchmark tests on four systems (Stardent's ST2000, Hewlett Packard's DN10000, Silicon Graphics' 4D 240 Power Series and Evans & Sutherland's ESV). They consisted of some computational performance tests (basic arithmetic operation, matrix calculations) and some application performance tests (two mesh generation programs, one multigrid mesh program, magnetohydrodynamic code, SOR - successive over-relaxation and the Navier-Stokes program).
JH was responsible for ensuring the benchmarked systems were installed. Three systems were loaned to the department for a period of at least a week. One of these (the Silicon Graphics) turned out to be large for a workstation and had to be sited in the computer room. A fourth system, the Stardent ST3000, could not be loaned to the department so JH benchmarked it at the Stardent UK office. The Stardent ST2000 was of course already here.
JH was also responsible for preparing and running graphics benchmarks. One, Xperf, tests the performance of the X11 interface. This is very comprehensive and is also a good test of reliability. Most systems failed some of the tests. The other, spheres, was written by JH and tests the advanced 3D capabilities of the systems. Unlike the other benchmarks, this needs careful porting when benchmarking a new system.
RM coordinated the report production, with RP providing substantial sections.A presentation on the work was given at EASE 90 in April. The project reported to CFTAG in April and June and the results of the assessment will be published in an appropriate technical series (RAL Report and/or EASE Technical Report).
JH has continued to manage the Stardent ST2000 and there is increasing use of the system for applications, mainly in CMD. Release 2.0, the first major software upgrade, was installed. When the X terminals were installed, work was needed on the ST2000 system to allow them to boot. Visual/3, a CFD Visualisation system from MIT was installed.
Further interactive devices, a spaceball and a set of dials were added to the ST2000.
The TITAN system was loaned to the ABACUS Unit at Strathclyde for them to mount application software and report on the suitability of the system for this purpose. The system is to move to Imperial College, where a CFD group will use it.
As a result of previous research an internal report is due to be issued. Before any other assessment work is continued, some feedback on specific requirements within the EASE programme is expected.
At the Parallel Computing 89 conference in Leiden, RP presented a paper Ray Tracing on the SCRIP Machine, which she wrote jointly with Dr Graham Megson. The paper was published in the conference proceedings book.
Although the ISO graphics standards GKS-3D and PHIGS are capable of expressing 3D concepts directly, they fall short when complex surfaces are required. PHIGS PLUS is an extension of PHIGS which allows polygonal and curved surfaces to be defined directly and also the lighting and shading that enables them to be viewed effectively. An ISO working group (ISO/IEC JTC1/SC24/WG2) is responsible for ensuring the progress of PHIGS PLUS towards being an ISO standard. In the process, technical improvements can be made. JRG is on the BSI and ISO groups responsible for PHIGS PLUS. Among the technical improvements are to make it more suitable for data visualisation and for use as the graphics system underlying postprocessors for STEP.
This is the first full reporting period for the ARGOSI project, which engages the attention of staff from all the divisions in Informatics.
Within CMD, the main activity of JRG and RM during the reporting period has been the Classification Workpackage, with the other Informatics project staff. The main purpose is to analyse applications which need graphics over networks and to analyse and classify their requirements, which may be expressed in terms of current standards or recommendations for new ones.
Some visits were undertaken, but in this reporting period, information has been gathered chiefly by means of a postal questionnaire designed by the workpackage, with the document being produced by Graham Reynolds of the University of East Anglia. The questionnaire was analysed at UL with software by DAD. Some preliminary results were presented at the April Project Review.
The Workpackage is due to end at end of September.
JRG is the Workpackage leader.
ERH continued his joint-appointment as Associate Lecturer in the Department of Electronic and Electrical Engineering at Surrey University. This involved contributions to the teaching in all three undergraduate years and at M. Sc level. Some lectures on pattern recognition in the final year Machine Intelligence Course were especially well received, being upheld as examples of lucid presentation in the Departmental Staff-Student Liaison Committee.
Towards this MSc, intensive work on the dissertation has taken place with the intention of submitting this early in the next reporting year.
Considering visualisation software, there are signs that flexible visualisation toolkits are beginning to emerge. In the coming year, the group will work with other groups with visualisation needs and in doing so understand where the new software falls short. Our initial impression is that some commercial systems could act as a basis for a useful visualisation system for RAL and EASE. The EMR at Swansea and the TITAN loan to Imperial College are opportunities for this. Where possible, we will work with other interested parties in UK academia, under the AGOCG umbrella.
It is planned that Mark 1 of IPAL will be completed and that it can begin to be exploited by interested parties in the EASE community.
Software for which the group is responsible will need to work in the EASE environment and thus occupy the group's attention.
The JFIT project Vision by Associative Reasoning will complete its first full year within the period and we will also be joined by an ERCIM fellow (Michal Haindl) during this time.
Contributions to standardisation of image processing will take place, participating in BSI and ISO work.
Assessments of hardware suitable for visualisation will continue to be needed.
The superworkstation equipment for which the group is responsible will continue to be managed and will be upgraded as needed.
All aspects of visualisation and imaging require significant resource and parallelism will need to be harnessed to an increasing degree in the future, depending on funding.
DCS is primarily a service Division. It serves the communities supported by the bodies which fund the Department. It also serves the other three applications Divisions within Informatics. Finally it makes its expertise available to the rest of RAL. Its main functions are:
As this report period ends the Open Days are rushing towards us. K F Hartley and F M Russell have been responsible for Informatics Department's contribution, mainly by maintaining interest and ensuring that those responsible for specific items - handouts, display boards and demonstrations - deliver on time. At the end of June it seemed unlikely that everything would be ready, but at the time of writing we now know that it was and was very successful.
The major innovation during this year has been our involvement in the British Computer Society's Professional Development Scheme. This has not been without its problems, particularly a late start and declining enthusiasm of the participants. It has also involved substantial extra work for the co-ordinator, supervisor and participants. There are indications that as we become more familiar with the scheme and the Industry Structure Model evolves the Scheme will meet the needs of IT staff, particularly new young members of staff and those wishing to take positive action to change the direction of their career development.
Reviewing the objectives for the DCSD in last year's Report makes somewhat dismal reading.
EASE+ funding has been secured for the present financial year but following a very successful series of presentations to ITAB's technical committees the Board reacted rather negatively and postponed a final decision until September.
The Alvey Programme is now finally completed but the files, and many others, have still not found their way into the Archive.
The funding situation is still difficult but not impossible. The period without SUN maintenance was not as bad as was feared and SUN were extremely co-operative about getting machines back onto maintenance. During this year the performance of IDUS under new improved releases of system software and the memory requirements of X have posed more serious problems. Our network is now sufficiently complex that it is very difficult to identify where the bottlenecks are with any confidence.
Informatics did participate fully, stylishly and effectively in the RAL Open Days, though they took place two weeks after the end of the report period.
The technology in H V Jenkins' office has not been updated mainly because no Department-wide decision has yet been made on the way forward. In fact HVJ left at the end of the year having completed her BTEC course and being recommended to someone starting up a new company by one C J Pavelin.
K F Hartley has been leading a working party to define the EASE Software Environment. A proposal and workplan were completed to schedule and accepted by CFTAG. The next phase, implementation, is likely to dominate the next twelve months.
KFH regularly attends the Inter-University Software Committee (IUSC) on behalf of EASE. This is one of a large collection of committees and working parties which bring together the views of HEI Computer Centres and present them to the Computer Board and CHEST. Several fruitful areas of common interest have developed between IUSC and EASE, including X terminals, algebraic software and finite element packages.
The highlights of the year must certainly include the organisation of EASE 90 on top of a full programme of Education and Awareness events. MSG staff worked extremely hard to arrange the event at relatively short notice, and are now hard at work on EASE 91. The initial' round of HEI visits is now 90% complete and most of the 700 plus questionnaires safely recorded in a database. For SSG the highlight was in February when no less than five members of the Group were all recognised as Achievers at one time. TSG has had a very successful year. GKS validation continues and progress is being made with CGM activities and the by-now ritual thanks and praise were given by CFTAG for further assessment reports. This should not obscure the fact that no two assessments are the same and that each requires considerable energy and ability from the team to achieve the high standard that everyone has come to expect. System Admin has worked heroically to keep IDUS running at all given the rate at which people and skills have been lost from the Section.
For the record it should be noted that this was the year of the final ending of our involvement with Prime Computers. Primes have been an integral part of our support of the engineering community for more than a decade. Their final passing was marked with due ceremony, bringing together many of those involved over the years. Truly this marked the end of an era.
Our scheme for ritually embarrassing members of the Division who have achieved over and above what might be expected has continued. The following have been honoured in this way during the past year:-
At the start of this period (October 1989), the group membership was:
Congratulations to SKC on gaining his Ph.D, and to ID on her wedding.
Congratulations also to RMK, AJL and WJH on their promotions.
JPB completed his six-month assignment with PLP and moved to SSG.
William Lui (KWL) joined in January to work with PLP on Graphics. RT transferred to Systems Support Group (SSG) at the end of February. Nicola Watkins (NJW) joined TSG from SSG at the same time, to work in Assessment. KML moved to Design Division in April.
At the beginning of June, the group membership was:
The Group has three sections: Assessment, Systems Administration and Text/Graphics.
The main objectives are as follows:
The Assessment Section began an evaluation of servers in October, with the intention of producing a report for the December meeting of CFTAG. Nineteen potential suppliers were identified and invited to take part by answering an Operational Requirement and providing a machine on loan to test in a network of up to 10 clients (of various makes). It was considered essential that machines should run Unix, fully support SUN's Network File System (NFS) and Yellow Pages, and be capable of connecting to the Joint Academic Network (JANET). Only four suppliers, DEC, MIPS, Solbourne and SUN agreed to take part; the main problem with the others seemed to be lack of Coloured Book software to connect directly to JANET. When the machines were benchmarked at RAL, it was found that only SUN was able to meet all the mandatory requirements.
With these disappointing results, the Report presented to CFTAG was accepted as interim, and the team were asked to extend the exercise (since several suppliers had reported that they would be able to take part in the near future). In the event, the only extra tests which could be undertaken concerned various other servers from the SUN range. CFTAG therefore decided to accept the report at this stage, but noted that it would be possible to provide a network with a connection to JANET by means other than via a central server. They have therefore asked for the evaluation to be repeated in the last quarter of 1990, but with the JANET mandatory requirement removed. Meanwhile, a public version of the report has been prepared and is being circulated (RAL-90-032).
In January, work began on the evaluation of the X terminal, the new peripheral which fits between the ASCII terminal and the workstation, giving a graphical interface instead of the earlier character based interaction. An initial difficulty was identifying the UK distributors, with a new name appearing almost weekly. The majority of these have had problems with supply, so it has proved quite difficult to arrange for the loan of machines. Thus it was possible only to provide an interim report to the April meeting of CFTAG. The final report covers seventeen machines, of which nine were benchmarked (although several turned out to be the same one under a different name!). The June CFTAG meeting approved three of the terminals assessed, but advised that X terminals were more likely to be of use attached to multiuser machines, rather than connected to a client/server network. A public version of the report is being prepared.
During this exercise, there has been a considerable interchange with the Inter-University Software Committee (IUSC) Working Party on Workstations, which is undertaking a similar exercise on PC-based X-terminal emulation software. It will be interesting to compare the results.
A second evaluation of 15 workstations was completed during the third quarter of 1989. A public version of this report was produced in October (RAL-89-1l0) and this has been widely circulated. A short article was provided in the Engineering Computing Newsletter (ECN), and several individual versions of the full report were generated for those Higher Educational Institutes (HEIs) engaged in tender exercises.
Following a request from CFTAG, the SUN3/80 was benchmarked further and, as a result, added to the approved list. Three other workstations have been benchmarked since the reports: an 8Mb Acorn with improved operating system, a Solbourne 4/500 (the first of the SUN clones) and the Data General AViiON/300 (not a new machine but having a new version of the operating system). The results will be included in the current exercise, which started in June. This will not cover all contenders, but will concentrate on the newcomers. It is noted that there has not been such a quantum leap in workstation power during this period (compared with the previous 12 months). However, there are strong indications that another significant increase is about to take place.
Improvements and enhancements to the benchmarking suite are under constant review. Several programs/packages in the Finite Element (FE) area have been tested on the various workstations available in ID and can now be incorporated in the benchmarking suite; these are LES (the linear equation solver from BIM2D), DFT, BERSAFE and UPSET (from SKC's thesis work). Work is almost complete on producing a Software Engineering benchmark based on ERIL.
A collection of GKS benchmarks has been gathered and made suitable for benchmarking. Successfully developed on SUN, they still have to be proved on a loaned machine.
A server accelerator board from Legato (called the Prestoserv) has been made available on loan from its UK distributor, Computer Marketing Services (CMS). This board improves throughput by increasing write speed and reducing processor overhead. Early results show that it does indeed speed up writes by over 200%, but the number of writes actually performed in our network is sufficiently small that the overall gain is much less. It is being considered as one of the possible ways of improving the Network File System (NFS) response.
ID, ASJ and occasionally RT have maintained the service by dumping the servers, restoring lost files, general housekeeping, maintaining printers and solving problems as they arose. ASJ took on most of the work of adding and deleting users, monitoring the use of filestore and reading system mail. Both ASJ and ID have been trying hard to persuade users to reduce their private filestore and making more use of project related filestore. ASJ has been updating the list of offenders for the Management Meeting (IMM), but it is hoped to solve all problems by friendly persuasion!
ASJ has been undertaking the usual security checks which seemed particularly worthwhile when they revealed some attempted security break-ins. but no success for the hackers. There have also been several cases of files being created in unusual directories, but so far nothing malicious has been found (see later comments on security).
ASJ organised the dump tapes so as to keep the right ones in the fire safe. ID prepared a plan for the room. and this has now been implemented. The servers are now in a horseshoe shape, making them much more easily accessible. ASJ arranged the disposal of further unwanted items: an Olivetti 3B2. racking and cables. She has also managed to make the 'storage area' look much tidier and smaller. Photographs have been taken for use on RAL Open Day Display boards.
ID and RT completed the upgrade of the servers. discless clients and some of the other disced Suns to release 4.0.3. The latest Fortran compiler was installed at the same time. ID has now upgraded all but one disked Sun to Release 4.0.3, and the work should be completed during the summer (currently delayed because of requirements for some software to remain at the previous version). The policy has been to make most disked Suns dataless clients with swap space on local disks; the exceptions being maple, elsie and laurel in Lab11 which will still be standalone machines.
ID installed the new Sun4 server, nfs1 (replacing an old SUN2) which is now serving 8 clients. Its filestore has been completely given over to the X11 software.
ID and RT moved the Yellow Pages master function from nfs2 to nfs3. RT,ID and ASJ altered the broadcast address throughout the local ethernet. working with NBMC. DSB prepared the NAG library for installation on Suns.
RT and ID planned and completed the merger of the Pyramid 98X and Workcenter (nfs4), with the aim to make nfs4 look almost like pyr-a. As usual ASJ was keeping everyone sane by keeping track of the many backup tapes involved in this transaction. The old pyr-a is being removed, and this should increase network reliability.
ID organised the removal of further obsolete equipment from Lab 11 and its further tidying. RT and ASJ took over Lab 10 from Technology Department and prepared plans for its eventual use, meanwhile making it available for the server assessment.
ID worked out a design for a permanent ramp in Lab 11 in collaboration with R18, which was to be in place by the end of March, but has now been vetoed by Health and Safety because there is insufficient room to provide a slope which adhered to accepted standards. Plans are now in hand for another more convenient temporary ramp.
ASJ and DSB have been helping to set up the training set of machines for demonstrations, especially with regard to technical advice on network connections. The demos have taken place in several locations around the country, as well as within RAL.
DSB receives and deals with about 250 queries a quarter, most of them straightforward, including reports of hardware faults. DSB made available documentation for users of the new Fortran compiler. DSB, KML and RT have been working on the distribution of software to other sites.
KML several times reviewed the deliveries to the Central Server sites, eventually providing most items which Sun had not supplied, with the exception of SunOS 4.0.3 upgrade tapes.
There have been requests from external sites for Ralpage, X25, spy, ten, and Coloured Books software to be placed on cartridge tape and distributed. DSB has actioned these requests. Tapes and manuals for Nag Mark 13 Fortran Library were produced and distributed to the remaining licence-holders. DSB organised this activity.
KML and RT passed lists to Sun of desktop machines to be removed from maintenance as an economy measure, and eventually agreed with Sun that the machines could be fully maintained again from 1 January 1990. Meanwhile ID, ASJ and DSB had to shuffle around the broken parts to keep as many systems as possible running. In the end 8 systems were unusable.
As last year, much effort was put into trying to reconcile the lists of maintained SUNs produced by us and SUN, in order to pay the maintenance bill. The problems caused by this each year are getting beyond a joke. It has therefore been decided to ease the situation by splitting the contract into three: Grant-related (given to Swindon), external to RAL (small number), and RAL site. This last includes non-Informatics machines, and we intend to go out to tender for this part. However, things have been delayed by the changeover from Harwell to RAL of the contracts team. The old contract (very long) exists on a word processor which RAL does not use, and it is impossible to transfer. Hence it has to be retyped. This has yet to be completed. A 6-month extension of the current contract has been requested to give more time to sort things out.
The Department Security Policy paper was at last produced, discussed at an open meeting, further refined by Division Heads and then issued. This has been a long involved process. However, the policy is now clear, and implementation notes have also been issued. The paper itself excited considerable comment, especially with regard to the Risk Analysis, and the Director requested copies be sent to all Department Heads in RAL and all Laboratory Directors in SERC! The next main job is to find time to explore the CRAMM software obtained from the Central Computer and Telecommunications Agency (CCTA), and also to plan for a security audit later this year. CRAMM attempts to automate the process of performing a Risk Analysis.
PDA is producing a Security Administrator's Guide, which will Cover all aspects of the requirement in the Department.
There were a couple of attempts by external people to break into the system. Neither was successful. One of the attempts was traced back to Switzerland. Others showed that various Universities had been successfully penetrated. We informed the managers.
The main preoccupation in the graphics area has been the preparation for the validation of RAL GKS by the National Computing Centre (NCC) testing service. This has involved running the NCC test suite with all of the RAL GKS workstation drivers, then locating and fixing bugs (both in the test suite and the GKS implementation). PLP and RMK have been working on this. Several update generations of the RAL GKS libraries were made in the process.
In addition, pick input has been added to RAL GKS. This now brings the implementation up to full level 2b functionality. Central Computing Department (CCD) had written general-purpose utility routines to handle pick input, and RMK incorporated these into the implementation on the Sun.
PLP tested the RAL GKS Computer Graphics Metafile (CGM) drivers (written by CCD) using the NCC test suite and produced an internal note on the results. He also revised the PostScript driver and fixed the problems discovered since its local release. This, together with a number of other bug fixes and new files, was then added into the 1.30 version of the RAL GKS Master Source.
PLP arranged the transfer of the latest RAL GKS Master Source across to CCD and provided assistance in the building of the UNICOS version of RAL GKS for the Cray.
At the request of Starlink staff, PLP arranged for the transfer to a DEC Station of all the sources necessary to generate the 1.20 release. This was later used for a successful build of the RAL GKS libraries on this machine.
JPB successfully ported GKSMVIEW, a program written by CCD for viewing GKS Annex E metafiles and CGMs, to our SUNs. Following this he made a number of improvements, always ensuring that machine independence was retained. Some new features were introduced, e.g. ability to switch input metafi1es, an additional output workstation (with a view to use the PostScript driver and thus provide a hardcopy facility) etc. JPB also enhanced the user interface, making the best possible use of the GKS input devices, and wrote two GKS Installation Project Notes, containing the necessary technical and user information. This concluded JPB's 6 month Trainee Scientist project.
RMK was involved in the setting up of the AGOCG Graphics Newsletter, and is on the Editorial Board for this. She also joined a working party looking at versions of GKS for use on PCs. The aim was for CHEST to negotiate a bulk deal for an implementation. This proved to be an interesting exercise as there isn't even a standard Fortran compiler in use in the academic community - CHEST have deals for three different ones and none of those allow cross-linking to C! So a side-effect of this work will be a new CHEST deal for a C compiler which will work with one of the Fortrans.
As part of her work for AGOCG, RMK visited University of Manchester Computer Graphics Unit to discuss the merge of their GKS-3D system with the current RAL GKS sources. (The Manchester GKS-3D was based on an early release of RAL GKS.) This project at Manchester is funded by the Computer Board through AGOCG.
KWL joined RAL in January and is working with PLP in the graphics team. He has successfully ported the RAL-CGM package to our Suns. This was written by CCD and allows the translation and viewing of files in all three CGM standard encodings. Having familiarised himself with the software as well as with the CGM standard and our computing environment, KWL produced the first user documentation of RAL-CGM. He then taught himself PostScript and has been developing a PostScript output driver for the same package. The backbone of the driver together with several basic output primitives have already been written and tested. This work on RAL-CGM is part of a larger project initiated by AGOCG to provide checking, translation and viewing tools for use with CGM files.
FRAH and DAD requested facilities for the translation of CGM files generated by SunPHIGS into PostScript. This was required for the slides/notes for a PHIGS tutorial which FRAH and DAD gave in June. PLP put together a set of utilities which handle this translation based around GKSMVIEW. This is an interim solution (as it limits the CGM to the elements which can be handled by GKS), and the utilities will be changed to use RAL-CGM once KWL's PostScript work is complete. However, the present version works well and has allowed FRAH and DAD to produce their tutorial material.
BDC joined the section in early September. He was involved initially in learning UNIX and related text processing systems such as FrameMaker and troff. During October and early November BDC arranged courses in FrameMaker and wrote and distributed course material. Four morning courses were held in collaboration with C. Kwong. BDC then took over user support for FrameMaker, including arranging for the provision of documentation for FrameMaker users. After gaining more experience of various aspects of FrameMaker, including using it with X-Windows, printing PostScript files and making Hypertext documents, he gave individual tutorials to six individuals who had been unable to attend the courses last year.
BDC installed FrameMaker Version 2.0 and configured it for use as an experimental service. He also wrote documentation on the configuration of the various versions and how to use version 2.0. BDC liaised with Contracts and our FrameMaker supplier over the purchase of an additional six licences. This now gives us a total of 16.
BDC arranged a UNIX for Secretaries course given by UMIST on 20 March, and assisted Systems Support in setting-up course facilities. He also wrote a review of the course material prepared by UMIST.
BDC was involved in studying the possible use of the ISO standard SGML, and SGML-based text formatting systems such as DAPHNE, with a view to serving the future text-processing needs of the Department. RMK and DAD have managed to obtain a copy of DAPHNE (for TEX!) and have been investigating its capabilities.
The question of using SGML was considered at a Workshop at Cosener's in March. This was organised by AGOCG, as there is a need for the distribution of graphics documentation in machine-readable form. RMK and DAD wrote a paper for this workshop, describing an attempt to mark-up part of the RAL GKS user documentation into SGML. RET and KFH attended this workshop.
Membership of the SGML User's Group was applied for.
RET is a member of the RAL SGML Working Party, which is looking at possible uses of SGML within the Laboratory. A report will be presented shortly to RAL Office Systems Panel (ROFSP), and this is likely to propose that the next RAL Annual Report be prepared using SGML.
RMK continued to handle the support and maintenance of DYB. The main achievement in this area was the (final!) solution of the problems with differing LaserWriter page offsets. The breakdown of the two remaining old-style LaserWriters resulted in Sun replacing them with new-style printers. This meant that the four service printers were all of the same type, and so (with much help from RT) RMK was able to install the necessary changes to DWB and TranScript to use the new-style page position. This ends the long-running page position saga!
There was a rash of troff queries in March - largely caused by several people preparing papers and foils for an ESPRIT review meeting!
In June, BDC started looking at printers, with a view to recommending a high-speed, double-sided postscript printer for use in the department.
The major task during the year has been preparing for the closure of the systems at RAL. Regular meetings have been held and PDA has been involved in these, advising and arranging for the tasks to be carried out in the time-scales available. The main area of concern has been the retrieval of data, especially from the Technology community (RLPC and RLPF). There was still some work outstanding at the time of shutdown of these machines (December) and it needed to be scheduled for action before the main Prime (RLPA) closed down.
In the end, more than 600 Mbytes of data were retrieved for users and then written to IBM formatted tape. On line data were transferred to other machines, sometimes using file transfer (FTP), but generally by writing tapes. The service on RLPA and RLPH closed on time and the machines were taken out of user service at the end of March.
RET continued to keep a check on the Department's finance state. It was apparent that it would be almost impossible for spend to keep within the set limits, since most of it was generated by existing commitments. One money-saving action was the removal of desktop Suns at RAL from maintenance for the last quarter of 1989.
During the first quarter of 1990, RET spent much time trying to monitor the Engineering Board (EB) spend. In order to do this more effectively, he learnt SQL on the IBM so that he could generate his own reports from the Finance Database. However, even this does not give enough information to be useful, since the various delays in the system mean that there is approx 4 weeks between a bill being paid and its appearance in the Financial Data System (FDS). By that time, it is too late to do anything. Attempts at persuading Bill Paying to send all invoices to the Department Head (as Associate Director for EB) for counter-signature failed, as they forgot to include the non-Informatics ones.
From April, there was a major change to all the project codes (now called Cost Centres) as a prelude to the introduction of MSA. MSA promises to provide more control, but it is not clear yet who will be allowed access. RET has spent much time generating Cost Centres for the Department, in the face of repeated policy changes from Finance (most of which involved starting from scratch again).
One of the major concerns this year has been the state of the filestore. Regular reports of the top 20 occupants were requested and Systems Administration were actioned to visit all those with holdings above 20Mb, and a "warning" level of 10Mb has been set also (for user filestore, not project).
A change request form for Lab 11 has been produced by ID. However, there is now little problem with space in Lab 11, following the various rearrangements.
The problems with performance have been discussed on a number of occasions. It is clear that more data is required to find out why the service sometimes appears to go very slowly. SSG have begun to prepare the necessary tools to undertake this task (see separate report).
A new plan for 1990/91 was prepared and this has been accepted with little change by Division Heads. In addition, plans for the introduction of an X service (ie provision of window systems based on X11) were discussed, but the need for at least 8Mb memory, and the large number of SUN 3/50s, has made planning difficult. Third-party memory has been considered, but this is difficult to install and there are potential problems with maintenance.
ID joined the Management team when RT was transferred.
The Grant Forecasting Model has progressed. The prototype has been tailored to accommodate all of the Boards. Tutorials have been held at RAL and PDA has assisted in these. During the period, representatives of Engineering, IT, ACME, Science, Bio, MSEC, Teaching Company and APS (all from Central Office), have attended these talks. The move to IBM started with investigations using SAS/SQL. PDA arranged for the transfer of the CPL,SPSSX and Fortran on RLPA to IBM code and this was completed on time (and validated). The grants database, QMF and SAS are being used to check out the model on the IBM in order that the boards at Central Office can assess the IBM version. The main emphasis is on Engineering, Science and MSEC boards at this stage.
As usual, CCD asked Informatics to run the three-week section of the Trainee Scientist course which was devoted to Unix. On this occasion, NJW acted as tutor. The training SUNs were set up in Atlas, and the trainees undertook a small project after having a course on Unix and C. The format followed the previously successful formula of dividing the trainees into groups, presenting them with a vague open-ended specification, telling them they had to produce a proper specification quickly, and then some programs and documentation to implement it. At the end of the course, RET and NJW ran "acceptance tests" to see how well the project met the specifications they had produced.
Over the years, the effective time available for this work has been eroded, due to pressure of time. As well as losing a day per week owing to day-release commitments, the period chosen started on a Bank Holiday! Everyone agreed that time was really too short to do the work justice. This has been passed back to the organisers.
The personnel of the Group is currently:
RT joined the Group in March 1990, transferring from Operations Section in TSG. KG changed from full to part-time working in February 1990, following the birth of her child.
Until March 1990 N J Watkins (NJW) also worked in the Group, in the area of protocol support. She then transferred to Assessment Section within TSG.
J P Barnett (JPB) joined the Group as a Trainee Programmer for the period January - June 1990.
Systems Support Group (SSG) has the remit to provide support for the systems and communications facilities used by the Department and the projects it undertakes. These facilities are in the areas of support for LAN based communications and for distributed systems services, for access to WAN based facilities, and for other systems services such as print serving.
The following Sections list the projects in which the Group has been involved over the period of this Report.
NBC continued to monitor developments with Yellow Pages from Pyramid. Further work on the RAL version was frozen as Pyramid's own version will be available at OSx 5.0.
NBC provided assistance to Assessment section during a Central Server evaluation. This included producing the operational requirement for manufacturers and providing technical help with benchmarking the machines for NFS performance. NJW assisted by providing functionality test suites for X.25 and Coloured Books on the machines.
IJJ has spent most of his time on projects involving high-speed networking. In particular, he has continued to work on the Remote Access to Supercomputer Project (RASP), which involves connecting Suns at Sheffield University to the Cray and IBM in Atlas. NBC and RAD were involved in the management of this project.
The work with Sheffield has involved the setting up of a 64 Kbit/s (KiloStream) private link between a Sun workstation situated in the Department of Mechanical Engineering, and a Sun fileserver at RAL. The fileserver has been connected to the Hyperchannel network, to which the Cray is also connected. Using this equipment it has been possible to provide the researchers at Sheffield with much improved access bandwidth to the Cray, and also to allow them to use forms of access (eg via NFS) not currently possible over JANET.
NBC has continued to liaise with the researchers at Sheffield and to arrange meetings, take minutes, order equipment and provide assistance to the people at Sheffield. NBC took up equipment to set up a small Ethernet segment in the Mechanical Engineering Dept. to connect together a Sun, Hitech and later a Sparcstation.
IJJ has continued to provide specialist support for this project, including visiting Sheffield in January to upgrade their machine from SunOS 3.5 to SunOS 4.0.3: this had been held off pending the availability of SunLink X.25 6.0 which was needed for SunOS 4.0.3. The researchers at Sheffield have found improvements in the turn-around times for their Cray batch jobs. IJJ is examining methods to improve the interactivity in the work cycle.
NBC has been collecting together information and reports to make up a project report on RASP. Major sections of the report have been written by RAD and IJJ, and NBC is now editing the report and filling in some of the gaps. The document will be passed to IJJ and RAD for further comments and additions.
NBC has been running some programs to obtain NFS statistics on the Sheffield machine for use in the report. He has also been running some further performance tests along the KiloStream line and will put these figures into the report too.
Access to the IBM and Cray at RAL is required by the Sheffield Sirius project and NBC, RAD and IJJ have been providing advice on equipment. NBC submitted a request to the EMEC Committee at Swindon and as part of a grant announcement money has been awarded for a software and memory upgrade on the Sun fileserver used at the RAL end of the link. NBC has been processing the purchases, which are: a copy of the hyperchannel driver - allowing the Sun Operating System to be upgraded to SunOS 4.0.3; copies of Sun's Internet Routing software; and a memory upgrade to 8 Mbytes for the Sun.
NBC provided advice to Prof. Hayhurst at Sheffield when funds became available for the purchase of another machine at Sheffield: a Sparcstation was selected. Dr Osman, Prof. Hayhurst's RA has asked for help in connecting the Sparcstation to the IBM. IJJ has helped in setting up IP routing from the Sparc to the IBM at RAL, and NBC tested FTP on the line.
RAD has monitored developments associated with the SERC Chairman's wish to see fast access to the Cray supercomputer provided to Engineering Board users in general. He has attended a number of meetings on this, and gave a talk at a meeting of grant holders interested in such facilities. This was based upon the technical possibilities available, and the experience gained through the Sheffield work.
IJJ (with RAD initially) has assisted CCD with the testing of special-purpose hardware known as IP Routers for high-speed networking over Wide Area Networks. This is an important technology that the JNT wish to promote so that effective use can be made of the upgrades to JANET, which are providing line speeds of up to 256 K bps. The testing has variously involved connecting IP Routers back-to-back through high-speed serial lines and connecting IP routers to Suns, in both cases directly and through X.25 exchanges, thus simulating the most likely hardware configurations to be employed.
Having been involved in the initial technical work, RAD subsequently set up an informal collaboration with CCD to investigate further the use of IP routing over X.25. It is likely that this work will be extended to include a link between Daresbury and RAL.
NBC and IJJ visited the Clinical Research Centre (CRC) which is part of the Medical Research Council (MRC) at Northwick Park Hospital, Harrow. They met with the head of computing services at CRC, Francis Rysavy and Geoff Gibbs of the same group and discussed shared interests in networking Suns over WANs and running services based on the TCP/IP protocols.
Overall the visit was very useful and stimulating. It identified a number of areas where it would be very valuable for SSG to collaborate with CRC CS.
Investigation of methods for Distributed Computing has been held up by the non-availability of the Network Computing System (NCS) software. NBC was unable to get hold of a copy of the NCS tape. Part of the problem was due to the upheaval arising from the takeover of Apollo by HP. Recently, IJJ has made contact with HP-Apollo and expects to receive a copy of NCS soon.
In connection with the RASP project, IJJ has started investigating the features provided by UNICOS, the version of UNIX running on the Atlas Cray.
SSG wish to investigate compute serving possibilities on the Cray and NBC put together a case for SSG to have time on UNICOS for testing connectivity etc. This was to be sent to CFC but the matter was raised at the RAL IP Group meeting who agreed that SSG should be allocated userids on the Cray.
NBC introduced JJSC to the Linotron systems on nfs1 and louis and attempted to liaise with Repro. and Scientific Admin. on developments. JJSC spent a lot of time getting to grips with the code for the Linotron, teaching himself C at the same time. There were several problems with both hardware and software which prevented successful spooling of documents to the Linotron. Most of these have now been fixed.
NBC and JJSC have been looking at the possibility of having a connection to the Neutron Division Ethernet village. A plan has been agreed and a Sun will be installed in R3, so that the system should be usable once the Ethernet link is enabled.
CCD now have a public domain network printserver which consists of lpd on a PC. NBC and JJSC investigated how this might be incorporated into a Linotron service. It has been decided that this is currently not a viable option.
A trial service will be started in August 1990.
NBC has obtained nhfsstone, an nfs assessment tool from Legato Systems Inc., and has been examining this software with a view to use in assessment. NBC has carried out some tests with a view to putting together a meaningful set of benchmarks to complement our existing NFS suite.
NBC discussed NFS testing with Assessment section and arranged with CMS Ltd. for a loan to carry out an evaluation of the Legato NFS accelerator board. NBC ran the nhfsstone benchmarks on Sun3 and Sun4 servers both with and without the accelerator board. The conclusion was that the board works as specified, but that it would not substantially improve the performance of the IDUS fi1eservers due to the nature of the work these perform. A report is being prepared.
There are plans to run NFS over WANs e.g. from Daresbury and this has brought the problem of site-wide UNIX uids into focus. NBC has been coordinating the policy on uids. Having talked to Operations and found that there was no firm policy NBC wrote a paper for the LAN management meeting on a policy for the distribution of user ids. NBC's paper was discussed at the first RIPG meeting at which there was general acceptance.
NBC has altered the newsclean script to deal with all the new news groups which have come in. He has also installed a patch to allow a quick start-up option which does not check for new groups. A bug with mailing to moderated newsgroups was also fixed.
For the first part of the period of the report NJW provided support for the IDUS Mail Service. This consisted of dealing with queries about the system, and of testing various new versions of X.25 and Coloured Books software released by Sun.
Since NJW's departure, JJSC has taken over general mail support and maintenance. All user queries to date have been successfully dealt with. This has involved getting to grips with the sendmail configuration files, liaising with the NRS people about the database and learning how to install a new configuration file on our mailserver.
JJSC has been involved with plans to move the mail system so that the current machine can be upgraded. This has required the installation of the new Sun X25 software (V6.0) and Coloured Books (2.0) onto another fileserver and testing it. There is still one bug in the X25 software that Sun are working on, but this will not affect the operation of the system when it is set up as the new (temporary) mailserver.
The first attempt to move mail between fileservers had to be abandoned when complications arose. JJSC has now discovered the source of the problem and a new attempt will be made soon.
RAD, KG, RT and NJW have spent time on a Pilot project set up to experiment with the provision of Directory Services in the UK academic community. Directory Services (often known as X.SOO) are a set of OSI services designed, among other things, to provide a human-accessible Directory Service to provide contact information for members of the community. This is seen as of particular importance in the area of electronic mail - with the expansion in the use of electronic mail it has become critically important that Directories exist to allow users to look up mail addresses of potential correspondents. In this sense X.SOO Directory Services perform a similar function as the telephone directories published by British telecom and other PTTs, along with their Directory Enquiries services.
Within the Pilot project there are a number of separate Directories connected together via OSI. KG and NJW set up a Sun system to provide the service at RAL, and KG installed and debugged the X.SOO software. RT now administers this system, and has written procedures in conjunction with KG to handle the data for the system. KG has developed a graphical user interface to the system, to allow Unix users to browse the Directory.
RAD runs the RAL part of the Pilot project, a joint undertaking between Informatics and CCD. He also defined a protocol for IBM users to access the Directory. This was implemented on the Directory Sun by NJW, and subsequently maintained and enhanced by RT.
NBC and APM were involved with changes to the IP addresses of machines on the ID Ethernet LAN. Tests were carried out prior to switching the whole network over. After further tests in which some problems were identified, NBC oversaw the change to a new IP broadcast convention.
NBC has been looking into ways of monitoring and improving performance on the ID Ethernet. He made out a case for a Trainee Programmer to write a set of utilities that would be useful. NBC wrote an IMM paper which formed the starting point for the work of the trainee - JPB. The latter developed a set of graphical tools to display in a convenient form the raw output of various utilities available in Unix to monitor system status. He also worked on a scheme to archive performance data for later use.
NBC has started to evaluate a copy of Sun's SunNet Manager.
NBC has been looking at ways of providing a modem connection to the ID LAN via one of the Bridge EtherPADs. APM set up one of the EtherPADS in the machine room and telephone lines have been installed in the Comms area. Demonstrations of modems which could provide a call back facility have been carried out. These should improve the security of the system, and also make reclaiming telephone charges a less complex task. NBC is now arranging for modems to be purchased and a dial-up service initiated.
JJSC, supervised by IJJ, has been trying to install NSE (Network Software Environment). JJSC studied the NSE documentation to get some understanding of the capabilities of the product. Installing NSE did not go as smoothly as desirable because of bugs in the install scripts. Problems existed in trying to determine what files would be modified when the system was installed on the YP master server - another Sun was used to try and trace the scripts as they were executed.
APM has compiled and installed a new version of tn3270 into lusr/ralfbin for Sun 3s. A compiled version for Sun 4s available in lu/publicfbin.sparc. APM has modified keymapping files and manual pages and has written shell scripts for use with the X Windowing system. APM has also written a paper comparing tn3270 with SIM3270 (a proprietary product that was tested earlier).
IJJ has installed compilers and other utilities from the GNU project of the Free Software Foundation. The initial impetus for this was to provide an up-to-date C++ compiler to be used by the X Toolkit evaluation team. These compilers have since proved to be useful to other users of IDUS and are now part of the service.
NBC has put up working Sun3 and Sun4 versions of mon.
NBC has written a script (rload) on nfs4 to provide the load average and number of users on Suns in Lab 11 designated as compute servers.
NBC wrote an IMM paper on evaluation of the Clearpoint memory boards which were installed in two Sun 3/50s.
The Department is considering the replacement of the Displaywriters used by the secretaries for word processing with more modern equipment that is integrated into IDUS. RAD is providing advice on the technical issues involved, and this may involve work by other members of the Group at a later stage.
APM has continued to act as the LAN Ethernet manager. As well as the regular tasks he has tidied up the cabling in the Human Factors Lab Video Room, made room on one of the EtherPADs for dial-in modems, and helped with Open Days planning.
NBC has been providing managerial effort for the X toolkit evaluation. This has included chairing progress meetings and writing the minutes. He has also looked at the availability of some commercial toolkits and obtained the ET++ distribution.
APM has been heavily involved with the X windows Toolkit Evaluation, and has had to learn to programme with Athena Widgets. APM started writing a tutorial on Athena Widgets and has been developing a test application as the focus of this work.
A user requirements questionnaire was sent out to sites and NBC typed in the data and produced a summary paper. APM then used this as the basis of a paper analysing the results in more detail and picking out the essential and desirable criteria.
NBC spent some time discussing the provision of an X service as part of IDUS with Design Division before ULM took up the issue (RAD has now produced a paper).
APM has installed OpenWindows version 2 (Beta) on a SparcStation.
APM has installed X.desktop on a Sun 3/60. Due to difficulties inherent in this program he has not yet been able to put it into a public area.
As part of his work under the ARGOSI project, RAD has contributed to work within the community on the production of a suitable mapping with which to run X Windows over OSI networks. This is important on two counts. In practical terms it is necessary to define a mapping that can be used over WANs such as JANET, which do not use TCP/IP. It is also necessary to have a viable OSI mapping to ensure that the forthcoming attempt to standardise X Windows within ISO stands a good chance of success.
RAD is secretary of a group set up by the JNT to co-ordinate activity in this area, and which acts as the technical advisory group to the relevant BSI committee.
ARGOSI (Applications Related Graphics and OSI Integration) is an ESPRIT II project concerned with the harmonisation of ISO Standards in the areas of graphics and OSI. It is aimed at identifying the needs of distributed graphical applications, and providing input to the Standards bodies where necessary to ensure that these needs are met. The project is also developing a demonstrator to show that Standards in these two areas can be successfully combined in a realistic application. RAD gave a talk about the project to the Department early in 1990.
RAD is RAL project manager for ARGOSI, and also contributes in a number of technical areas on the OSI side. He has participated in the Classification Workpackage which aims to classify real applications according to their graphics and OSI requirements. He was Workpackage leader for the Connectivity Workpackage, which provided an initial estimate of the feasibility of using international PDNs as part of the demonstration. He has also contributed to work on providing an OSI mapping for X Windows (see above), and on the development of a functional profile for FTAM to support the handling of Computer Graphics Metafiles (CGMs).
KG has mounted and tested the ISODE FTAM system. She is also working on the enhancement of this to enable it to be used in the demonstration.
RAD has continued to carry out this contract which provides the academic community with technical advice and recommendations on the suitability of various Ethernet bridge products as ways of linking LANs. The work has consisted of gathering information from suppliers, and in carrying out in-depth tests of products. Regular summary documents are published within the community based on these tests.
This work is scheduled to end in October 1990, due to the increasing involvement of the Group in new areas of work.
The Group has been given effort to provide communications support for this Programme, based at Birmingham University. So far it has been unclear exactly what the Programme's requirements are. RAD has been monitoring events, and RT and NBC have made a number of visits to HEIs involved in the Programme to enquire about the facilities they require.
The Design Management Committee is setting up a series of EDCs at HEIs throughout the country. RAD provided the Committee with a technical assessment of the communications aspects of this project, and has subsequently provided advice on the hardware and communications aspects of HEI bids.
NBC has made some changes to the Postscript business card programs for the DAO and has produced various logos.
RT provided a mail responder to enable readers to obtain electronic copies of the text of Engineering Computing News, which was handed over to Management Services Group.
KG and RT installed the ibrowse interface to the X.500 Directory on a JNT Sun for the RAL Open Days.
The BCS PDS started on January 1st this year. NBC is a supervisor and participant. IJJ, APM, RT and KG and JJSC are participants. All have had to produce career development plans. RAD is a supervisor only.
NBC took part in 4 HEI visits which were organised by D Lomas.
RT has taken part in 3 HEI visits organised by G A Lambert.
RAD has taken part in 2 HEI visits organised by J W T Smith.
The following are the main reports of a technical nature written during this period by members of the Group.
The objectives of the Group continue to provide those services for Informatics Department which it is sensible to centra1ise. Some are provided on behalf of the whole Department, ego The majority of the EASE Services, The Departmental Administrative Officer, others for specific projects, such as ESPRIT 962.
Sheila Davidson (SGD) remains responsible for the production of the Engineering Computing Newsletter. This has now developed into a routine monthly event with everyone involved in the production cycle working, the majority of the time, to an agreed timetable.
The original objective of a significant increase in the circulation has been achieved. Well over 5000 copies are distributed each month. The range of articles has increased with the major portion originating from members of the user community. Obtaining articles is a time consuming activity but SGD has become highly proficient at chasing up any hint of a good story. The concept of regular monthly items from special interest groups is now firmly established, but plans to update the presentation are taking longer than was originally anticipated.
Susan Hilton (SCH) was diverted from her original tasks to commence producing the Graphics Newsletter. This is now also produced monthly, in parallel, with the ECN, and after the initial teething problems were resolved production of this is now handled by Janet Smith (JRS).
Production of the EASE glossy has proved to be a long winded process with (SCH) being diverted onto other more pressing activities whenever progress was made. The major difficulty was obtaining any agreement on the form this product should take. At long last CFC agreed on a general format and production of the initial draft is now taking shape.
Revision of the EASE schedules was passed to a special committee and again now that specific responsibilities have been defined some real progress on the task has been made. It is intended that this document will also be available to the community over the network and the layout has been designed specifically to facilitate this.
Education and Awareness has been defined as one of the high priority areas of the EASE programme. In December 1989 Professor Nicholson, late of the Control Engineering Department of Sheffield University, agreed to take on the role of Education Co-ordinator. After three changes to his Contractual terms of reference, when he became in turn, Education Co-ordinator, EASE Co-ordinator, then EASE Monitor, he too took the path of his predecessor and resigned. This adds confirmation to the fact that:
GAL assumed this responsibility again and to date has not retired, yet!
The major activity of the Education and Awareness team (GAL, JWTS, David Lomas (UMIST) plus invaluable assistance from SCH) has been the visits to the Engineering Departments of all HEI's in the UK. One objective of this activity was to attempt to measure the impact and effectiveness of the EASE policy. The major instrument in obtaining this information was a questionnaire, designed by the committee, which proved to be a long iterative process which ended up satisfying nobody. The real problem was that it was never made clear what information was required, and worst of all, when obtained, how it was to be analysed.
However, regardless of the problems the majority of the sites have been visited, with over 700 interviews taken place covering some 75 HEI's. Some analysis of the data has already taken place, and in part has been acted upon, but if the good will generated by the initial visit is to be capitalised upon, follow up activity is essential. This will form a major activity during the coming months.
E & A events form another significant activity of the Group and everyone ie.(GAL, SGD, SCH, JRS, JWTS, and David Lomas) have all made contributions. Events include:
| Function | Event | Date | Location | Attendance |
|---|---|---|---|---|
| Tutorial | SUN System Admin | 17.8.89 | RAL | 6 |
| Workshop | UIMS | 18.9.89 | UMIST | 85 |
| Workshop | OOPS | 19.9.89 | UMIST | 70 |
| Tutorial | SUN O/S Introduction | 20-21.9.89 | UMIST | 12 |
| Tutorial | SUN O/S Introduction | 22.10.89 | UMIST | 10 |
| Tutorial | SUN System Admin | 31.10.89 | RAL | 8 |
| Seminar | How to Support Long Lived Software | 9.11.89 | RAL | 80 |
| Tutorial | SUN O/S Introduction | 15.11.89 | Leicester | 14 |
| Tutorial | SUN System Admin | 16.11.89 | Leicester | 8 |
| Seminar | AI Techniques for Engineers | 6.12.89 | RAL | 90 |
| Workshop | SGML | 5-7.3.90 | RAL | 35 |
| Tutorial | Introduction to UNIX | 10.3.90 | RAL | 10 |
| Seminar | CFD Community Club | 22.3.90 | RAL | 60 |
| Tutorial | Introduction to SUN O/S | 28-29.3.90 | RAL | 12 |
| Seminar | Computer Algebra | 3.4.90 | Sheffield | 70 |
| Tutorial | Introduction to SUN and UNIX | 4-5.4.90 | Sheffield | 14 |
| Seminar | Integrated Knowledge Based and Conventional Systems | 11.5.90 | Edinburgh | 40 |
| Workshop | Introduction to PHIGS | 29-30.5.90 | RAL | 25 |
| Seminar | SGML | 5.6.90 | Manchester | 50 |
| Tutorial | Introduction to SUN O/S and UNIX | 19-20.6.90 | RAL | 6 |
| Tutorial | Introduction to SUN O/S and UNIX | 25-26.06.90 | RAL | 6 |
The event of the year deserves special mention - EASE 90 which was held at UMIST from 18-20 April. This was attended by some 280 people over three days, included 58 speakers plus a large exhibition, was put together in less than 4 months, and by general agreement was an outstanding success. This would not have been possible without the wholehearted efforts of Susan Hilton, Janet Smith and Sheila Davidson who all deserve special recognition for all their hard work. It is now intended that there will continue to be an annual EASE event and plans for EASE 91 to be held in Birmingham on 25-27 March are now well advanced.
GAL has visited all Server sites prior to installation of the equipment to discuss the terms of the contract and resolve any problems. Installation was successfully completed at Essex. Leeds, Strathclyde, Bangor and Bradford Universities; Bath were allocated a Server during the last round but installation has been postponed due to lack of finance.
All sites were offered places on the System Administration Courses and almost without exception these have been accepted. This highlights the value of this activity especially as the course content is specifically targeted towards the needs of our community, and, important to all, the course is free. It is worth noting that the value of our System Admin Courses has spread beyond those sites where CFC funded Servers have been located. Nearly 50% of the course attendees have come from Institutions running SUN Servers paid for from Computer Board, other grant, or Institution funds.
GAL acted as Secretary to the Computing Facilities Technical Advisory Group (CFTAG) until December 1989. KFH, RET and GAL were members of the ECFE until it was disbanded in June. Even with the demise of these Committees the level of follow-up action continues undiminished.
GAL provided management support to the project. The project was a contributor to the ESPRIT Conference and Exhibition in Brussels in November 1989, having both an exhibit and hosting a workshop.
The project's final review was held in June 1990 but due to the failure of two of the partners to complete aspects of their individual contributions, a further final review is scheduled for November 1990.
As usual the project has necessitated the production of vast amounts of paper, the technical reports and the financial and manpower reports. All made inroads to the loss of the rain forests.
GAL was responsible for the rundown, shutdown end disposal of all the Informatics Department Prime facilities. Pat Athawes was responsible for all the day to day activity and achieved the almost impossible task of retrieving the vast amount of data which had been archived during the 12 years of the ICF. Further details are given in 5.2.25.
Final shutdown was marked in the traditional manner with a wake attended by some 50+ people who had made various contributions to the success of the project. The final switch-off was performed by Professor A MacFarlane, who had been Chairman for the ICF Programme Committee.
The first cycle of the scheme was completed at the end of June, with 16 members of the Division participating. KFH remains the Co-ordinator of the Informatics participants, with Professor Hopgood as the Super Co-ordinator and GAL acting as Secretary of the organising Committee.
Carol Barnes has continued to provide a day to day administrative service for all staff in the Department. The constant battle to keep ahead of all the Department's accommodation moves remains a feature of her work.
At least the level of the workload in the DAO's office has been (LAO) has been recognised and Lynne Oldfield a welcome addition to the section, CB remained Secretary of the Informatics Division Heads Meeting, and of SERC Suggestions Scheme Local Awards Committee.
This work predominated between June and December. PJH developed a facility meeting a wide range of departmental requirements. The aim of this project is to maintain information about departmental contacts. Subject interests, committee memberships and newsletter distribution requirements are recorded. Two large mailing lists were issued to establish the initial data. Facilities include ability to display lists of users, to view details of any user, and to define subsets within the list. It is possible to produce individually addressed letters to members of the subset, and typed address labels for every member of the subset. Data can be created, edited or removed. There is the facility to make selections of users, such as all Transputer users in the North West.
Some maintenance work was necessary chiefly to circumvent a bug in a new operating system. Opportunity was taken to improve the layout of reports after attending a course on Postscript.
This project developed programs written in a mixture of Fortran and REXX on the IBM CMS service and C and Postscript on the IDUS service.
PJH attends CCSUM (Central Computers Site Users Meeting) and now represents the whole of the Informatics Department. The structure of accounts used by the department had not been changed for many years and no longer represented the structure of the department's work. A new structure has been worked out which will be implemented during the summer of 1990.
This work began in October. PJH works for the Deputy Director as the Resource Officer for Rutherford Research Services.
The work of the Resource Officer for RRS has three aspects: bids, monitoring and approval. The work on bids is for the RAR allocations and Forward Looks. Monitoring covers the use of resources, out turn and surpluses. Approval covers advice on new projects and reviewing the RRS programme. The job requires liaison with Finance Division, Contracts, Departmental Resource Officers and Programme Managers. PJH attends the laboratory finance subcommittee.
As well as general responsibility for all EASE contracts GAL has been specifically responsible, with RET, for supervising the contract at UMIST. This is now coming to its end and has proved of considerable value to the project.
EASE 90, UMIST, April 1990, GAL, SGD, JRS
During the report period the staff of ITDU remained constant
However, the name of the Unit changed to reflect the changes that took place in the Department of Trade and Industry in April, 1990. For the historical record it is worth noting that the Alvey Programme was run by the Alvey Directorate in Millbank Tower. This was jointly manned by SERC and DTI staff, together with industrial secondees. Towards the end of its life its headquarters were moved to Kingsgate House, Victoria Street, still as the Alvey Directorate but now firmly part of the DTI. This was seen as an anomaly by the DTI Mandarins and so it was renamed the Information Engineering Directorate (IED). RAL staff supporting IED were therefore placed in the IED Unit. The latest change was when it became IT Division (ITD); note the significant change from Directorate to Division. This consolidated the process of absorbing Alvey into the DTI. The IED Unit was therefore renamed the ITD Unit. This has caused a certain amount of confusion as the SERC body responsible for IT matters is, and always has been, the IT Directorate (ITD). These arcane name changes do represent a significant difference of opinion - SERC sees IT as very much more of a directed programme than DTI.
The objective of the IED Unit is to provide a home in Informatics Department for Informatics Department staff who spend much of their time at Kingsgate House, and whose duties do not fit naturally into the structure of Informatics Department.
The main objectives set for Mike Russell this year have been achieved. He completed the selection and appointment of Monitoring Officers for all the projects in the first and second calls. Many of these projects are now running and the MOs are producing their reports on project performance. A list of potential monitoring officers has also been generated for possible future use.
The reports produced by the MOs are used to judge the performance of the projects. These reports are analysed by Mike Russell for signs of possible project difficulties and he reports any such indicators to the Director of Systems Engineering of the ITEAP in the SERC/DTI IT programme. All projects are defined in terms of deliverables and these are set out in the Level 2 plan of the project. A major initial task of the MO is to ensure that a satisfactory Level 2 plan is received from the Project Officer. Delays in producing this plan are an early sign of possible problems. The production of deliverables is an excellent indicator for monitoring project performance but the MOs have to use their skills to monitor the success of team work and collaboration between partners in the projects.
Reports by the MOs for the remaining projects in the Alvey programme also are included in the coordination analysis made by Mike Russell. To ensure that the Mas set acceptable standards for the projects two one-day seminars were organized by the RAL ITD unit. These seminars covered many aspects of monitoring and included inputs from experienced Mas. As a result of the first seminar the guide book on monitoring prepared by the DTI was altered to reflect the special problems of monitoring IT projects.
Tony Lubbock has continued to work for Adrian Wheldon, who is now based at Swindon, on helping to manage the External Interfaces and Special Systems parts of the Systems Engineering components of Joint Framework for Information Technology programme. There continues to be some activity generated by remaining Alvey IKBS and Systems Architectures projects, particularly managing extensions to these projects. For new projects AJKL is mainly concerned with preparing the final proposals and ensuring that the necessary contracts are let as quickly as possible.
Individual research has been conducted by Mike Russell into solid-state recording processes. This work has been funded on a 50% basis by RAL during the past year. Earlier work suggested that the recording processes involved two distinct mechanisms. One depended on inelastic scattering to produce ionization which then caused a proportionate decoration of the track by an impurity present in the mica. Detailed study of the tracks has confirmed that a second process is involved that is independent of ionization. This is a major finding and calls for a new approach to track recording. In this new process the energy for modifying the lattice to record the track comes from relaxation of the lattice itself rather than from the particle.
This new process has assisted in explaining many of the lines in mica which previously were not understood but had been shown not to be caused by charged leptons. These lines are consistent with the propagation of energy and momentum through the lattice in a coordinated way by solitons or supra ballistic phonons(SBPs). To test this hypothesis an experiment was devised by Mike Russell and is currently under way in collaboration with a group from Frankfurt. The energy in the solitons or SBPs is obtained from an incident energetic particle, by relaxation of stored energy in the lattice or by combinations of these.
A collaboration to test the ionization dependent recording process in an artificial mica in the laboratory was set up with a group in Shanghai. The troubles in China have caused this work to be halted.
The second recording process is consistent with the triggering of a phase transition in a polyphasic material. Although the initiation of a phase transition in liquid/vapour systems by charged particles is known, evidence for this process in the solid-state is new. Such a process may allow for external control over phase transitions in any polyphasic material, including metallic alloys. Action has been taken to gain patent protection for this process. A laboratory test of this process should be attempted and is now under consideration.
Mike Russell was asked to take on the early work on co-ordination of Informatics Department's contribution to the RAL Open Days, which are to take place just after the end of the Report Period. He attended all the early meetings of the RAL Open Days Panel and instigated the actions necessary for the production of display material of all kinds, as well as helping to define the overall shape of the Open Days. Later on he was helped by KFH and most Informatics Staff in what was a very successful venture.
IEATP Conference, Southampton, March, AJKL
Decorated track recording mechanisms in Muscovite mica. FMR
