Contact us Heritage collections Image license terms
HOME ACL ACD ICF SUS DCS G&A STARLINK Literature
Further reading □ Overview1. February 19772. April 19773. June 19774. October 19775. December 19776. April 19787. July 19788. February 19799. June 197910. October 197911. January 198012. April 198013. November 1980
C&A INF CCD CISD Archives Contact us Heritage archives Image license terms

Search

   
ACDLiteratureRapid Response
ACDLiteratureRapid Response
ACL ACD C&A INF CCD CISD Archives
Further reading

Overview
1. February 1977
2. April 1977
3. June 1977
4. October 1977
5. December 1977
6. April 1978
7. July 1978
8. February 1979
9. June 1979
10. October 1979
11. January 1980
12. April 1980
13. November 1980

No 10 October 1979

1. CONTROL ENGINEERING SOFTWARE

The June meeting of the Interactive Computing Facilities Committee approved a proposal by the Control Engineering Special Interest Group. The proposal consisted of a programme of software development to support control engineering users of the ICF. The following report is from Dr M J Denham of Kingston Polytechnic, the SIG chairman.

The programme, which was fully discussed at the Control Engineering users meeting in April 1979, consists of the following major activities:

  1. establishment of a control subroutine library
  2. mounting one or more simulation packages to provide a comprehensive dynamic simulation facility
  3. support for the UMIST and Cambridge multivariable design packages on the ICF
  4. developing and mounting an existing or new data analysis and system identification package.

This programme covers three years during which the work will be carried out by two applications programmers based at Kingston Polytechnic. Work will start as soon as suitable staff can be found; interested readers should contact me for details of the posts. A small team from within the SIG will manage the project.

The control subroutine library is planned as a major tool for the development of control design software in the next decade. It will cover the full range of algorithms currently used in controller design packages, implemented with numerically efficient and stable computational methods. Expert advice on this aspect is being sought from numerical analysts in an associated research programme for which SRC support will shortly be requested. With the high standard of documentation and detailed definition of standard interfaces proposed. We hope that the resultant library will prove both easy to use and very flexible in aiding the development of new control design methods while minimising software development costs.

Simulation is a major component of research into control system design methods. The SIG will take the initiative to make available a comprehensive simulation facility on ICF machines. A detailed assessment of user requirements and an analysis of available packages will be carried out in collaboration with the UMIST user group studying the provision of DEC-10 simulation facilities. Eventually a widely applicable facility will be provided covering interactive input and analysis of data, continuous time and discrete event simulation. The user will be able to link it with his own subroutines for system optimisation.

There will be no support for developing new packages for multivariable control system design. However, the existing implementation of the UMIST and Cambridge design packages on the Facility will be consolidated. User support, program maintenance and documentation will be improved or extended where necessary.

Also planned is the development and implementation on the ICF of a comprehensive data analysis, parameter estimation and system identification package. The work programme includes an analysis of existing packages; a new package will be implemented in conjunction with Professor Douce's group at Warwick incorporating the major facilities of these packages for man/machine interaction, data storage and data analysis. In the next few months, I hope to visit all the institutions where control engineering user groups exist.

2. SOLID STATE DEVICE FABRICATION

SRC has allocated £2 million to set up improved facilities for academic research workers to fabricate integrated circuits ("silicon chips") and other microelectronic devices. Existing facilities at four universities - Southampton, Edinburgh, Surrey and Sheffield - have been augmented in a programme coordinated through the Solid State Devices (SSD) sub-committee of SRC's Information Engineering Committee. The coordinator, Bill Turner (Technology Division, Rutherford) will handle enquiries from potential users, who should contact him at an early stage if they want to use the facilities. This work is not part of the ICF but has links with it. Southampton and Edinburgh will carry out silicon processing, Surrey ion implantation and Sheffield the epitaxial growth of III-V compounds. An Electron Beam Lithography Facility has been set up at the Rutherford Laboratory to produce masks for the fabrication process. Use of the facilities will normally be through SRC grants; a streamlined process ("continuous stream") is available for small grants. A system of "pump priming" has also been introduced, similar to that already in use in the ICF. A newsletter called "Micro fabrication" has been produced to keep the research community informed about developments. Anyone who wants to receive copies should contact the Editor in Technology Division, Rutherford Laboratory.

THE ELECTRON BEAM LITHOGRAPHY FACILITY

This facility was formally inaugurated by the Rt Hon Shirley Williams on 16 July 1979. It is based on a Cambridge Instruments EBMF-2 electron beam microfabricator which will generate masks from which integrated circuits will be produced. Minimum line width and separation in the masks is currently used at 2 microns but the machine is capable of resolving to a micron and below. Chrome mask manufacturing and copying processes are available.

COMPUTER-AIDED MASK SPECIFICATION

The standard input format to the EBMF-2 is that produced by the GAELIC suite of programs (see "Rapid Response" No.6, section 4), which is available on the Edinburgh DEC-10 and the PRIME 400. Files are transferred over links to a PDP-11 which controls the electron beam. A library of subroutines is available for incorporation into the user's own program and the package is being interfaced to other existing programs. GAELIC uses geometrical representations of silicon structures. Work is under way to supplement this with an interface to a data-base containing topological details of the specifications, and to enhance GAELIC's graphics facilities.

3. ELECTRONIC SYSTEMS SOFTWARE

Following the acceptance by the Interactive Computing Facilities Committee of the proposals from the two Special Interest Groups for Electric Circuit Design (ECD) software, those SIGS were disbanded. A Digital and Analogue Circuit Design (DACD) Steering Committee has been formed; it will monitor the development, mounting and support of software for electronic system design and analysis on ICF machines. The first meeting was held on July 19th; the members are:

The DACD group at Rutherford is providing CAD software tools; this is a large project with limited in-house resources. An outline of the system and some component packages now exist on the ICF; much work remains to be done and some specific contracts will be placed with R&D groups outside Rutherford throughout the duration of the project. We wish to identify University groups willing and able to undertake commercial contracts (Extra-Mural Development Contracts) to mount various items of ECD software. In addition it would be helpful if we could obtain information about the available machine-readable files on ECD data. Any research workers able to help in this way are asked to contact Peter Dewar who heads the DACD team.

We intend to keep potential users of our software in touch by means of the DACD Bulletin. The first issue will appear shortly. It will consist of short chapters giving details of the circuit design software available on all ICF machines, with references to other manuals and short introductions to the system software on each machine. Future issues will either provide additional chapters, worked examples. new applications etc, or replace out-of-date issues. If you would like a copy please contact Ian Benest.

4. DATABASES ETC ON PRIME

PRIME DBMS

The PRIME Database Management System (DBMS) is an implementation of the CODASYL Data Base Task Group 1971 report, with extensions. It is available on the ICF PRIME 400 at the Rutherford Laboratory. It provides generalised facilities for creating and maintaining structured data bases, with benefits such as data independence, reduced data redundancy and less programming effort to maintain and develop application programs. As a 'CODASYL-type' DBMS it allows users to represent both hierarchical and network-type data structures.

A DBMS is useful for large amounts of structured information, where there are various basic types of data and they are interconnected. The user can define the types of records and data items in them and specify the relationships required between different record types. The DBMS then maintains the relationships and allows the user to access and update the data, using the logical structure without being concerned about the physical structure underlying it. Where a body of data is shared by a number of users, a DBMS reduces the problems associated with updating multiple copies. PRIME DBMS uses subschemas to allow users to have different "views" of a data base and to look at only the part required for one application.

Application programs using PRIME DBMS on the ICF are written in FORTRAN with extra statements to access the database. There is currently no stand-alone query language for PRIME DBMS.

PRIME DBMS allows concurrent access by multiple readers and/or updaters and incorporates various security features. It contains a comprehensive logging and recovery system and utilities for saving and restoring databases. The facilities are still being investigated; More details are available from Mattie Keane or Ann Walter.

MIDAS ON PRIME

MIDAS is a data management system for setting up and maintaining large indexed sequential (IS) files. These contain records which may be accessed either in order or individually via an index holding keys and pointers. A MIDAS file may have up to 20 indexes (standard IS methods usually allow only one) and the keys can be of various types including integer and character. Access to the data is through subroutine calls from FORTRAN and utilities allow the database to be set up easily. MIDAS is useful where a large amount of data is required but the structure of the information is not complex enough to warrant the use of a full DBMS. An introductory guide has been published as PRIME USER 25.

STATUS

STATUS is a full text Information Retrieval system developed at AERE Harwell, and has been purchased for the PRIME at Rutherford. It offers efficient on-line retrieval of data, macros and updating. The basic division of text, ie a record, is defined as an 'article'. Articles can be grouped into 'chapters', which can be used to narrow the searches. Each article can be divided into named sections to restrict the search even further. The response to any search is the number of articles that satisfy the question. The complete article, or title only, can then be displayed for all or a selection of the retrieved articles. The simplest search is for articles containing a single word, truncated word or phrase, and more complicated searches can be made using logical operators.

Examples are:

Q PHYSICS?
finds all articles containing the word 'PHYSICS'
Q PHYS*?
finds all articles with the root 'PHYS' such as PHYSICS. PHYSICAL, etc.
Q PHYSICS + COMPUTER SCIENCE?
finds all articles with the word 'PHYSICS' and the phrase 'COMPUTER SCIENCE'

A Thesaurus and Report Generator are being developed at Harwell. There are currently 3 data bases on the PRIME for demonstration of STATUS, including the PRIME manual. More information is available from Judy Lay at Rutherford.

5. DEC-10 NEWS

HARDWARE CHANGES AT UMIST

Swapping on the UMIST DEC-10 has been exclusively to an elderly RC10/RD10 controller/disk system. The 2.5 Mbyte disc, although purchased from DEC, is in fact a Burroughs product and has been increasingly hard to maintain, partly because of difficulty in obtaining spare parts. In June the ICF Committee agreed to its replacement by an RHP06 controller/disk system, still on a separate channel. Delivery is expected later this year, but meanwhile the RD10, obviously sensing that the end was near, quietly laid down and died. This seriously degraded the UMIST service during August/September.

The new disk drive has 200 Mbytes capacity and during installation will be interchanged with one of the existing RP04 drives, so releasing another 100 Mbytes for user files while providing 100 Mbytes for swapping. A decision on replacing four blocks of MA-10 memory (vintage 1970) is expected soon.

OTHER NEWS

The DEC-10 network, which uses DECnet proprietary networking protocols, has been operating very reliably over the last 4 months. It consists of a basic triangular trunk network linking the Edinburgh and Manchester nodes to a DN82 nodal processor at Rutherford Laboratory and is also connected to Dundee University's DEC-10 KL. The ICF will not make major enhancements to this network; future communications between SRC and University computers should be based on the newer X-25 (and higher level) protocols with which DECnet is incompatible. Work is progressing well at York University on the development of a gateway between any DEC-10 (or DECnet) and X-25; in the long term this will be used to link DEC-10s to the rest of the SRC network. Meanwhile a few additional connections are still being made using DECnet. One is between the UMIST DEC-10 and the VAX 11/780 at Leeds in which ICF has a part share. The link should enter service later this year.

6. PRIME COMPUTERS - PROGRESS

There are now four ICF PRIME computers: the dual 400 at Rutherford, and the two upgraded 400s at Nottingham and East Anglia. Nottingham has been linked via PRIMENET, using an X-25/binary-synchronous connection, to Rutherford for several months and the sites now run common systems; all software distribution is over the link. East Anglia, whose upgrade was only recently completed, still awaits provision of the line to Rutherford. Two further ICF PRIME Systems (550's which have superseded the 400) have been ordered recently for Sussex University (School of Engineering and Applied Sciences) and City University (Dept of System Sciences}. The latter was financed by SRC's Control Engineering Committee but will be run as an ICF machine.

A User Committee has been set up to represent the interests of users of the ICF PRIME computers. Its terms of reference are similar to those of the DEC-10 User Committees (see below). It met at the Rutherford Laboratory on 27th September; the membership is as follows:

A one-day introductory course is being prepared for current and prospective users of the PRIME. It may be repeated at new PRIME sites.

7. GEC COMPUTERS - PROGRESS

The 5 ICF GEC 4000 series machines have been performing well apart from some disc failures at some sites, and a bug (now cured) in the micro code of the plotters. A new version of the operating system is being released. This is a half-way stage to GEC's version 3 and entails a change in the library, so all code has to be re-linked. Mrs Jackie Hutchinson is now working full-time on user support for the GEC machines.

A meeting of the managers and potential managers was held in September: this proved very useful and many of the ideas put forward are being implemented. The meeting also gave an opportunity for the various managers to meet and chew over their experiences and problems. A further meeting will be held around the turn of the year. The next machine to be installed will be at Bradford in November.

SRC NETWORK PROGRESS

The network code in the GEC 4000 series machines is now complete. It is possible to send files between the GEC computers and to access them interactively from each other; the file transfer system should soon be available on the IBM 360/195. The main use of this facility has been to improve the support of the machines, ranging from answering complaints to delivering new software. It is now easy to update software, including operating systems, without travelling to sites. This will improve the productivity of SRC staff and the speed with which faults can be put right. Also, it is now far easier for support staff to help users. The HASP access to the 360/195 is now via the network and is much more reliable than the earlier version of HASP.

The network code, written by Dave Toll, Andrew Dunn, Graham Robinson and Tony Salter, has been remarkably robust and economical. The project has absorbed roughly 2 man-years and consists of 80-odd Kbytes of code and data. The SRC batch network has had steady expansion with the addition of further network workstations at some sites. There has been a steady move towards the various 'standard' protocols, with minor changes here and there which have had little effect on users.

8. USER COMMITTEES

Regular flow of information between users and the management was seen as important when the ICF was set up. The Rutherford Laboratory has therefore set up three User Committees, one each for the DEC-10s and one for the PRIMEs. A GEC 4000 User Committee is also being set up.

The chairmen of the three existing committees are members of the ICFC: they and the other members are invited to serve for two years. The committee members are a representative group either by virtue of their extensive use of facilities or as delegates for a number of smaller users. The chairman of the DEC-10 committees are: Edinburgh Dr A D Milne (Edinburgh University) and UMIST - Mr O S Mills (Manchester University). At the meetings, the relevant installation managers are required to be present as ex-officio members of the Committees.

Meetings are held three or four times per year. In addition to regular reporting of installation performance, usage and current developments in the ICF, users can bring to the notice of the ICF matters for discussion either to improve the service or to satisfy new user requirements. Recently the Committees at Edinburgh and UMIST have been presenting the user view in the discussions about the future of the DEC-10 services. As the installations of ICF Computers are completed the maintenance of good services will become a dominant part of the ICF work, and contact with the user view will be even more necessary. Users should therefore see the Committees as offering a direct means of influencing the development of the Interactive Computing Facility which seeks to encourage user participation in its affairs.

9. OTHER NEWS

CAAD SIG

The Special Interest Group for Computer-Aided Architectural Design organised a seminar on the use of databases and database management systems in CAAD. It was held at the University of Bristol on 21 September. Professor C M Eastham of Carnegie-Mellon University reviewed DBMS development in the USA and described the latest development in the GLIDE building development system. Mr E M Hoskins of Applied Research of Cambridge gave a tutorial on the BOS database management system which is being bought by the ICF, and a demonstration was provided. Short talks were given on INGRES, STATUS (see above) and AESOP - the last two were also demonstrated. We are grateful to Bob Phillips and the University of Bristol for the work they put in to make the seminar a success.

10. FINALLY

Observant readers will detect a minor change in the address below. This is the result of the merger of the two computing Divisions of the Rutherford Laboratory, as announced in the last issue.

Peter Smith, Computing Division, Rutherford Laboratory, Chilton
⇑ Top of page
© Chilton Computing and UKRI Science and Technology Facilities Council webmaster@chilton-computing.org.uk
Our thanks to UKRI Science and Technology Facilities Council for hosting this site