Minutes of the Meeting of the BCS Fortran Specialist Group
held at BCS, 13 Mansfield Street, London on 6 February 1992
Present: Tim Bartle Salford Software
Carol Hewlett LSE
Peter Holland SSL
Ramesh Krishna Met Office
Chris Lazou University of London
Zohair Maany NAG
Mike Nunn CCTA
Mike Roth AEA Technology
Nick Saville Private Consultant
Neil Smith DRA Bedford
Paul Wardle Met Office
John Wilson Leicester University
John Young PE-MOD
Due to the late arrival of the Chairman, Mike Roth, the
Vice-Chairman, Chris Lazou, took the chair. He opened the
meeting by welcoming everyone and thanking them for coming.
2. APOLOGIES FOR ABSENCE
Apologies for absence were received from Martin Counihan,
Mike Geary, E Golton, David Muxworthy, John Reid, Lawrie
Schonfelder and Dave Vallance.
3. MINUTES OF THE MEETING OF 16 MAY 1991
The Secretary, John Young, explained to the meeting the
reasons for the minutes being presented out of order and
apologised for not being able to complete the minutes of the last
meeting of 28 November 1991 in time.
4. MATTERS ARISING
The problems over the mailing for Salford Software Marketing
Ltd still had not been satisfactorily resolved. The
questionnaire which was sent to organisations and firms enquiring
what their future plans were for Fortran 90 compilers,
introduction of Fortran 90 features, Fortran 90 tools etc in the
autumn had been completed. Three replies had been received and
the results were printed in Computing on 6 February 1992. A copy
of the text appears in Appendix A.
5. REPORT FROM X3J3 REPRESENTATIVE
Regrettably, no X3J3 representative was present at the
meeting. It was noted that Miles Ellis who normally made the
report had recently changed his job. It was suggested that,
perhaps, a report could be sent to the meeting if Miles was
unable to attend. The Secretary, John Young, reviewed Miles'
report given to the meeting on 28 November 1991 and it was agreed
that little had happened since then. It was not known whether
the ANSI Ballot had been completed and whether there was now an
American Fortran 90 Standard.
David Muxworthy sent a brief report on WG5 activities. He
requested that people wanting to attend the WG5 in July in
Victoria, British Columbia get in touch with him. John Wilson
said that he had no further news.
6. THE DRAFT CONSTITUTION FOR THE GROUP
A copy of the draft constitution for the Group had been
circulated to members. The Chairman, Mike Roth, opened the
discussion by asking the meeting whether, firstly, the Group
wanted its own constitution, and secondly, were there any major
changes to be made to the details. After a short general
discussion the meeting agreed that it would be useful to have a
constitution for the Group and then proceeded to detailed
Mike had appended a series of questions to the draft and
these were dealt with by the meeting. It was agreed that the
Treasurer was responsible for the Group's finances and should be
submitting financial statements to the BCS. It was also felt
that the Treasurer should be responsible for the collection of
annual subscriptions from the Group's non-BCS members. The
representative on the Specialist Groups Management Committee
should be the Chairman with power to send to a substitute. It
was agreed that presently a Management Committee was not
necessary as this was the function of the Group's ordinary
The proposal to create extra posts was not accepted by the
meeting as the jobs were currently done by the existing
committee. All officers are to resign and seek re-election at
each AGM. The Chairman is responsible for the policies and plans
of the Group and for reporting its activities. There was a query
on who are corporate or sustaining members but it was thought
that these would have the same rights as ordinary members. It
was thought that the number of members necessary to call an EGM
should be 7, the same number as required for a quorum at an AGM.
There was also a discussion on who should be signatories to
cheques on the Group's accounts and it was agreed that the
current rules would apply. [Secretary's note: See minutes of the
meeting of 25 October 1990].
It was noted on the draft constitution that the numbering
of the items was not sequential. The Chairman, Mike Roth, said
he thought this was due to the computer system used to create they
document. He would ensure that the numbering was correct on the
revised document to be presented at the AGM. The major
correction to be made was to item 8h. The first sentence should
read "Every proposal at an AGM shall be decided by a majority of
the members present".
7. BCS BUSINESS
Specialist Groups Management Committee
The Secretary, John Young, had attended the December SGMC
meeting which proved rather acrimonious and vociferous with the
larger specialist groups leading the way. The new Chairman,
Brian Oakley, apologised to members for what had happened at BCS
citing the recession, the computer purchase and the HQ move to
Swindon being the main reasons for the BCS problems. He
confirmed that the bank charges would not be borne by specialist
groups. There was going to be a real problem on finding meeting
rooms and it was suggested that groups should get together. The
Chairman then announced that there was to be a special meeting
of the specialist groups on the current financial problems. It
was also announced that the service level agreements were
postponed. BCS HQ required that it holds all specialist groups'
membership lists but it had yet to work out the necessary
The meeting then elected Russell Winder of HCI and John
Mitchell of Computer Audit to be the specialist groups management
committee's representatives on the BCS Council.
The main item of the meeting was a report from A1ine
Cumming, Head of Education, on Information Systems Engineering.
The item on Membership Grades was postponed. Several new
specialist groups were then proposed. A questionnaire to
specialist groups about their external links was distributed.
Annual General Meeting
The Secretary, John Young, had attended the BCS Annual
General Meeting on 9 January. He reported on the main items of
business which had already been circulated to members of BCS.
The major items which are likely to affect the Group were the
staff reductions, the loss of 13 Mansfield Street for meetings
and the freezing of the Gold Account. The problems of Chartered
Engineer status was also discussed.
Request from Rome
An organisation in Italy thought to be connected with the
EEC had written to the Chairman, Mike Roth, about names and
addresses of Fortran Users in the UK. The meeting agreed that
the Group's Membership List was not relevant and that a reply was
to be sent that such a list did not exist.
The Vice-Chairman, Chris Lazou, reported that he had agreed
to write an Industry Report on Fortran & C and development tools
by Unicom Seminars. The meeting thought that this request had
already been passed around a number of members of the Group.
Chris asked the meeting whether the Group would be prepared to
sponsor the Report which involved just adding its name with no
money involved. The benefits included members' discount, peer
review payments and 2: user survey. The meeting agreed, in
principle, to this request provided a disclaimer was included.
Any suggestions to be sent direct to Chris.
The Secretary, John Young, had received a request for
information about Training Courses in Fortran. The Group offered
the names and addresses of several companies which could provide
the kind of tuition that was required.
Whilst no member of the Group could receive payment for
working for the Fortran Group the meeting felt that legitimate
expenses, in particular, travelling costs, could be paid for from
8. DATES OF MEETINGS
It was confirmed that the Annual General Meeting of the
Group would be held on 14 May 1992. The Secretary, John Young,
was standing down and it was also probable that the Chairman,
Mike Roth, would not be standing again. Nomination forms would
be sent out at the appropriate time. The Secretary was asked to
arrange a venue and a speaker with an IBM flavour.
The dates of the next four meetings were arranged as
14 May 1992 Venue yet to be arranged
Annual General Meeting
3 September 1992 Blacknest near Reading
Visit to Seismic Data Centre
26 November 1992 Venue to be arranged
25 February 1993 Venue to be arranged
9. DATA STRUCTURES IN FORTRAN 90
The meeting proceeded with an informal discussion on Data
Structures in Fortran 90. It was regretted that there were no
formal speakers on this subject so the meeting was asked to
contribute in any way possible. A summary of the discussion is
included as Appendix B.
7 April 1992
The Fortran Specialist Group of the BCS has carried out a
survey on behalf of the BCS and various interested parties. This
survey welcomed any comments on plans for developing standard
conforming Fortran 90 products. It received the following
The Numerical Algorithms Group claims to have produced the
world's first Fortran 90 compiler.
NAGware Fortran 90 compiler - a full ISO standard
NAGware Fortran 90 Test Suite - a comprehensive set of
Fortran subroutines to test compilers on every aspect of
the Fortran 90 ISO standard;
NAGware Fortran 90 Tools - a set of programmers' tools
including pretty printer, spaghetti unraveller, cross
reference tables, declaration standardiser, portability
verifier and a precision transformer;
NAGware Fortran 90 Library - a library of mathematical and
statistical routines in Fortran 90.
Edinburgh Portable Compilers offers a Fortran 90 compiler,
which will be available in the fourth quarter of 1992, and also
an interactive debugger, which will be extended to operate with
the Fortran 90 compiler when it is released.
Microsoft Fortran Professional Development System version
5.1, which has been shipping in the UK since July 1991, supports
a number of Fortran 90 extensions.
Microsoft is planning full support for Fortran 90 in future
products, but does not expect to see this until the end of 1992.
Programming Research plans to extend QA Fortran, its Fortran
88 static analyser which performs complexity, portability and
reliability checks to cover the Fortran 90 standard.
What is a "Data Structure" in Fortran 90?
The meeting agreed that a Data Structure is formed by
defining one or more derived types by a 'structure constructor'.
This Data Structure would have a name which would be treated like
any other variable in a Fortran program.
The preferred way to define a Data Structure.
A Data Structure can be defined in any subprogram and
referenced in the standard manner. However, the meeting agreed
that if it was defined in a Module with variables defined as
Private then the Data Structure cannot be corrupted and its use
checked at every reference. It was also noted that Structures
of Data Structures could be defined.
Using Data Structures.
The meeting agreed that Data Structures had to be one of the
major improvements in Fortran 90 giving the software designer a
very powerful tool. Its use makes COMMON, EQUIVALENCE, and
INCLUDE redundant. Members present were keen on the use of Data