Minutes of the Meeting of the BCS Fortran Specialist Group

    held at the University of Manchester's Nuffield Radio

   Astronomy Laboratories, Jodrell Bank, near Macclesfield

on 5 September 1991

Present:        Miles Ellis          Oxford University

                David Lomas          University of Salford

                Mike Nunn            CCTA

                Mike Roth            AEA Technology

                Lindsey Savage       MCC

                Lawrie Schonfelder   University of Liverpool

                David Shone          University of Manchester

                Neil Smith           Defence Research Agency, Bedford

                Julian Tilbury       University of Salford

                John Wilson          Leicester University

                John Young           PE-MOD

Also present from NRAL Jodrell Bank:

                David Shone          Tom Muxlow

                Simon Garrington     Roger Noble

                Lindy Brett          Michael Garrett

                Neal Jackson         Phil Powell

                Euston Luckle

l.      WELCOME

        The Group was welcomed by David Shone who had organised the

visit to Jodrell Bank. The Chairman, Mike Roth, warmly thanked

David for finding the time to organise the day which was much

appreciated by the members present.


        Apologies for absence were received from David Muxworthy and

Les Russell.


        Even though there were two X3J3/ISO representatives present,

it was decided that Miles Ellis would report on the recent WG5

meeting in Lund and the X3J3 meeting in Nashua.

        The WG5 meeting in Lund at the end of June was the first

meeting after completion of the Fortran 90 Standard processing

and was best described as a "wary" meeting.

        A number of items were discussed which had little to do with

current standard work. There was a discussion of ancillary

standards such as parallel processing, Fortran bindings for POSIX

and LCAS, and CLID and CLIP. The meeting agreed on what to do

next in Fortran and also discussed who owns Fortran, the US or

the rest of the world. Both WG5 and X3J3 will become involved

in interpretations of the Fortran 90 Standard and ISO will allow

certain revisions. However in the US, once an ANSI standard is

accepted, no changes can be made to the standard. Hence there

could be revisions to the international standard which would not

be applied to the US standard. Also the US will have two Fortran

standards; FORTRAN 77 will be an archival standard and Fortran

90 will be another standard. Also the US cannot adopt an ISO

standard whereas other countries, like the UK, can.

        The ISO Fortran 90 Standard, which is identical to the ANSI

standard, was published at the beginning of August. The ANSI

Fortran 90 Standard had to go to a ballot of X3 which is probably

completed. However, there will be a number of NO votes but

probably not enough not to pass. There will be no revision of

this single standard. The ANSI Fortran 90 Standard will probably

be dated 1992.

        At the Lund meeting a small sub-group with representatives

from various countries was set up. This group was given the

mandate to develop procedures for handling any revisions of the

Fortran 90 Standard and to advise on how Fortran will be

developed in the future. Currently, WG5 does the planning with

X3J3 doing the work, so work items can be different where they

should be the same. One possibility is for WG5 to operate as

X3J3 does now.

        The X3J3 meeting at Nashua in August, like the WG5 meeting

at Lund, did little work on the Fortran 90 Standard. The

principal objectives of the meeting were to address the procedure

for requests for interpretations and clarifications of Fortran

90, to draft responses to three protest letters, to respond to

the WG5 Lund resolutions, and, like the WG5 meeting, to review

collateral standards. X3J3 decided to start the new round of

interpretations with a "clean slate" by dropping all outstanding

requests. The meeting approved the WG5 Lund resolutions and

concurred with the establishment of the WG5 sub-committee for

looking at how the next Fortran Standard would be produced.

        The X3J3 also had a tutorial on the Journal of Development.

This Journal initially would preserve appendices from previous

drafts of the Fortran 90 Standard and could also be used as a

forum for new ideas in Fortran. A volunteer had been found for

a potential editor and an ad hoc committee had been formed.



        The Chairman, Mike Roth, had prepared a list of questions

about the use of Fortran and, in particular, Fortran 90 to be

sent to firms and groups. The SGMC had approved the form, in

principle, but needed to see the final version. The meeting

discussed the draft copy and a number of changes were proposed.

        Computer Journal Algorithms Supplement

David Muxworthy had written to Mike Roth concerning the

latest issue of the Journal in August 1991 containing a final set

of Fortran algorithms. The Group agreed with David's comments

and agreed that a letter to the editor of the Journal voicing the

Group's concern would be sent. The Secretary was to write and

send an appropriate letter.


        After an excellent buffet lunch, the Group were joined by

some members of staff at Jodrell Bank for a talk on "The Fortran

90 Standard" given by Lawrie Schonfelder of the University of

Liverpool Computer Laboratory. This talk was followed by an

"Introduction to Jodrell Bank" given by Tom Muxlow who gave a

short history of Jodrell Bank and described some of the latest

work being done. The talks were followed by an informal tour of

the facilities at Jodrell Bank led by David Shone.

John Young


31 October 1991