Minutes of the meeting of the BCS Fortran
Specialist Group held at the Rugby Club of London,
49 Hallam Street, London on 14 May 1992
Present: Tim Bartle Salford Software
Miles Ellis Oxford University
Mike Geary NAG
E Golton RAL
Carol Hewlett LSE
Peter Holland SSL
Chris Lazou University of London
Mike Nunn CCTA
Mike Roth UKAEA
Martin Stanworth Taywood Engineering Ltd
John Young PE-MOD
1 APOLOGIES FOR ABSENCE
Apologies for absence were received from David Muxworthy,
Les Russell and Dave Vallance.
2 MINUTES OF THE MEETING OF 28.11.91
The minutes of the meeting of 28 November 1991 were
approved. The Secretary, John Young, apologised for the delay
in publishing the minutes.
3 MATTERS ARISING
The letter to the editor of the Computer Journal Algorithms
Supplement concerning the last issue containing Fortran programs
nad still not been sent. The meeting felt that it was probably
too late and not worth sending the letter. [Secretary's note:
An enquiry concerning the letter has since been received and an
appropriate response has been made].
Membership reminders had finally been sent out by the
Secretary. A total of £80 had been contributed to the Group's
funds.
The inclusion of two items from Salford Software Marketing
Ltd in the last two mailings has completed the Group's
obligations to Salford. It was noted that the Group did not
receive a written response from the BCS concerning this matter.
A copy of the Fortran 90 Questionnaire which was sent to
groups and firms was not included in the minutes. It was agreed
a copy should be included with the master copy of the minutes.
4 MINUTES OF THE MEETING OF 6.2.92.
Miles Ellis requested that his name be added to the
Apologies for Absence in the minutes of meeting of 6 February
1992 which were then approved.
5 MATTERS ARISING
It was reported that Jim Bartle, Miles Ellis, David
Muxworthy and Mike Roth would be attending the WG5 meeting in
July in Victoria, British Columbia.
The meeting felt that Appendix B on "Data Structures" was
inaccurate in some respects and wrong in others. The following
points were made. A "structure constructor" was for the
definition of a constant and really nothing to do with Data
Structures. A Data Structure was defined by a "Declare"
statement which was also valid in Modules.
A question was raised of where to obtain a copy of the
Fortran 90 Standard. It was thought that a copy of the ISO
Fortran 90 Standard could be obtained from the BSI at a cost of
circa £80.
6 REPORT FROM X3J3 REPRESENTATIVE
The X3J3 Representative, Miles Ellis, reported to the
meeting on X3J3 activities over the past eight months. The last
X3J3 meeting had been in November with a meeting number of 120!
The next X3J3 meeting was due to take place in Indiana from 25
to 29 May.
In the US the Fortran 90 Standard had finally been passed
unanimously by X3, with Boeing being persuaded to change its vote
to YES. Fortran 90 was due to become the ANSI standard in the
very near future.
Miles went on to present various points which are detailed
in Appendix A.
7 REPORT ON PARALLEL PROCESSING ACTIVITIES
The Vice-Chairman, Chris Lazou, reported on his activities
in Parallel Processing. He was now Chairman on the European
Committee which had set up various committees and sub-committees.
There was now a complete proposal for a Fortran 90 binding. With
the advent of Multiple Parallel Processors there were certain
areas that needed to be addressed, like data distribution, so the
binding had not yet gone through the standardisation process.
There followed a general discussion on Parallel Processing
including a reminder to members that there is a Parallel
Processing Specialist Group.
8 THE CONSTITUTION FOR THE GROUP
The Chairman, Mike Roth, presented the corrected
Constitution for the Group. Copies were passed round at the
meeting with an apology for not being able to include it in the
recent mailing.
There was some discussion on item 6d on signatories to the
bank accounts but it was agreed to keep the original wording.
The meeting unanimously agreed to accept the presented
document as the Constitution of the Fortran Specialist Group but
the formal vote was delayed until the afternoon and the AGM.
9 ANNUAL GENERAL MEETING
(i) Apologies for Absence
Apologies for absence were received from David
Muxworthy, Les Russell and Dave Vallance.
The minutes of the last Annual General Meeting were
presented and approved.
(iii) Matters Arising
The proposal to spend less time on business and more
on technical matters had not, so far, been implemented.
(iv) Chairman's Report
The Chairman, Mike Roth, presented his report which is
reproduced as Appendix B. The Group held 4 meetings during
the year and published 2 newsletters.
(v) Treasurer's Report
The Treasurer, E Golton, presented his draft accounts.
He guided the meeting through the figures which were still
provisional as he was not sure of some of the BCS HQ
figures. He reported that after two years in the job and
with the tightening of financial controls at BCS HQ the
finances of the Group were now under stricter control. If
HQ made a transfer, now he at least, was notified. The
final accounts are presented as Appendix C.
(vi) Acceptance of the Constitution for the Group
Following the discussion on the Constitution for the
Group in the morning session it was proposed and accepted
unanimously on a formal vote that the document presented
becomes the Constitution for the Fortran Specialist Group
of the British Computer Society.
(vii) Election of Officers
The retiring Chairman, Mike Roth, remained in the
Chair for the nomination of the new Committee. The
retiring Secretary, John Young, had received only one
nomination for the post of Chairman; Chris Lazou being
proposed by Mike Nunn and seconded by Les Russell. There
were no other nominations so Chris Lazou was duly elected
to the post of Chairman.
For the post of Vice Chairman, John Young was proposed
by Chris Lazou and seconded by Miles Ellis. There were no
other nominations so John Young was duly elected to the
post of Vice Chairman.
There were no nominations for the post of Secretary in
spite of pleas in a Newsletter and on the Nomination Form.
As there was no-one in the meeting prepared to stand for
the post of Secretary it was agreed that the post would
have to remain vacant. The retiring Secretary, John Young,
agreed to be Acting Secretary for the next two meetings.
The Treasurer, E Golton, was prepared to stand again
and was proposed by Mike Nunn and seconded by Chris Lazou.
There being no other nominations, E Golton was duly elected
to the post of Treasurer.
(viii) BCS Business
The new Chairman, Chris Lazou, announced that he
intended to seek membership of the British Computer Society
which he had let lapse. Under the new Constitution the
Chairman is required to be a member of the BCS. Miles
Ellis and Mike Roth agreed to sponsor Chris in his
application.
The meeting endorsed a vote of thanks to the retiring
Chairman, Mike Roth, for his year's hard work, in
particular, in drawing up the new Constitution. The
meeting hoped that he might be able to attend some of the
Group's future meetings. There was also votes of thanks to
the retiring Secretary, John Young and to the retiring Vice
Chairman, Chris Lazou, both of whom remain on the
Committee.
Nobody from the Group had been able to attend the
Specialist Groups Management Committee meeting held on 18
March 1992. As no minutes had been received there was
nothing to report. A letter had been received from the
Chairman of the SGMC, Brian Oakley, requesting the Group's
views on the BCS Membership Structure.
(ix) Any Other Business
The Secretary, John Young, had received a letter from
IBM inviting attendance at a Technical Computing Update on
20 May 1992. No-one at the meeting was able to attend.
The Group had been invited to attend the Software
Development 92 Show at Wembley in the User Group Lounge. It
was agreed that a rota would be drawn up for the 3 days,
16-18 June 1992.
The meeting approved the inclusion of the notice of
the meeting on Fortran 90 at the Rutherford Appleton
Laboratory on 9 June 1992. The Chairman, Vice-Chairman and
the Treasurer hoped to attend. A request for the inclusion
of flier by Wiley Publications in a future mailing was
approved.
Miles Ellis suggested that the Group should publicise
its activities in Computing, Computer Weekly etc. He also
suggested that the Group might consider publishing its own
newsletter under the title of "Fortran Matters".
It was reported that some American Universities were
rumoured to be dropping teaching PASCAL and taking up
Fortran 90.
The following dates were agreed for the next four
meetings of the Group.
Thursday, 3 September, 1992 - Blacknest
Thursday, 26 November, 1992 - IBM
Thursday, 25 February, 1993
Thursday, 13 May, 1993, AGM
No venue has been arranged for any 1993 meetings.
10 PAST, PRESENT AND FUTURE FORTRAN
No speakers were forthcoming to talk on this wide ranging
subject so the meeting held an informal discussion.
It was agreed that the three Fortran Forums held in the
1980's were excellent in informing the Fortran Community of the
new features and the progress of what was originally called
Fortran 8x and finally called Fortran 90. Perhaps the Group
should consider organising another Forum now that Fortran 90 was
standard.
The meeting identified two future roles for the Group. The
first was to publicise and make aware that Fortran 90 was a
powerful, modern, general computer language especially with
compilers becoming available. The second was to set the
requirements for future Fortran, particularly as input to WG5,
over the next 2-3 years.
The format of the Group's meetings should also be re-
considered with perhaps a presentation in the morning and
business in the afternoon. The Group should also consider the
advantage on the use of publications. Miles Ellis offered to
produce a discussion paper.
John B Young
Acting Secretary
30 July 1992
[Minutes for the AGM part of this meeting were also issued as a
separate document for use at the 1992 AGM. They are appended to these
minutes.]
Report on X3J3 activities in recent months
The last meeting of X3J3 (meeting 120) was held in Milpitas, California, from 11th to 14th
November, 1991. Meeting 121 will be held in Terre Haute, Indiana, from 25th to 29th May,
1992. Meeting 122 will be held in Seattle, Washington, from 3rd to 7th August, 1992,
immediately following the WC-5 meeting which is being held in Victoria, British Columbia,
from 27th to 31st Iuly, 1992. I attended meeting 120, and will attend the WG5 meeting and
meeting 122; I am unable to attend meeting 121.
At the end of meeting 120, Jeanne Adams resigned as Chair, and Jerry Wagener resigned as
Vice-Chair. Jerry was subsequently appointed to the Chair by X3, and Maureen Hoffert (of
Hewlett Packard) was appointed as Vice-Chair.
During the meeting Henry Katz announced that he, working with the Center for Standards
of the Department of Defense, wished to announce that they were recommending to the DoD
that the Secretary of Defense reject Fortran 90. His statement continued "This is based upon
a consensus reached with the major DoD Laboratories and Centers. Our plan is to stay with
FORTRAN 77, include all of the bindings developed for FORTRAN 77 and include MIL STD
g 1753 and certain other vendor extensions plus some of the neater things in Fortran 90. Our
view is that those who fell in love with Fortran 90 can use ADA." Henry has always been
probably the hardest of all the committed NO voters, and it is not clear whether this
recommendation is likely to be adopted by the DoD or not. Watch this space.
Apart from these two matters, meeting 120 was almost totally concerned with "interpretation
processing". The procedure agreed by X3J3 for this work is that an interpretation must be
considered, and approved, at two consecutive meetings before it is formally accepted. At the
first meeting the problem is discussed in subgroup and a proposed interpretation brought
to the full committee. This interpretation, or a modified version, must be passed by a simple
majority of those at the meeting. The provisional interpretation is then published in the pre-
meeting papers for the next meeting, at which it must achieve a two-thirds roll-call majority
to be formally approved.
Fifteen interpretations had been tentatively approved at meeting 119 in Nashua, New
Hampshire, last August, and these were all formally approved during meeting 120. A
further 31 tentative interpretations were approved during meeting 120 and will be roll-call
balloted during meeting 121. All accepted interpretations will be collected in a new standing
document, X3J3/S20, the contents of which will be publicised from time to time through such
means as publication in the ACM Fortran Forum.
During the period since meeting 120, the email airwaves have been very busy with a number
of other interpretation issues, and these informally discussed issues will be dealt with
formally during meeting 121.
The large number of requests for interpretations, and the embarrassingly large number of
errors already identified in the Standard, have led to a strong feeling amongst X3J3 members
that the future evolution of Fortran should proceed in two phases. The first phase should
be directed at producing a "maintenance release" in 2-3 years' time, consisting of a revision
of the Fortran 90 Standard to correct errors and incorporate interpretations; it would include
no new features. The next major revision, which would include new features based on public
demand, would be produced on a somewhat longer timescale. X3J3 voted overwhelmingly
(21-0-1) in favour of a maintenance release being published in about three years, and only
slightly less overwhelmingly (18-1-3) in favour of the maintenance release and the next major
revision being developed concurrently.
X3J3 strongly believes that it, and only it, should be responsible for the detailed development
work for the next Fortran Standard(s), but approves the direction currently being taken by
the WG5 L12 Ad Hoc Group to propose that WG5 should develop a set of formal
requirements and then delegate the detailed development to some other (national) body. On
a straw vote on the subject of whether a distributed development model should be used, with
separate parts of the work being delegated. to different national bodies, X3J3 voted 0
(exclusively) - 0 (heavily) - 18 (sparingly) -2 (not at all) - 1 (undecided).
While I do not support the idea of distributed development, which could only work with far
stronger central management controls than are feasible in the international standards world,
I do not believe that the problems between the U.S. and the rest of the world, which reached
a peak in 1988-89, will be avoided next time without much stronger control over X3J3 (or any
other development body) by WG5.
As a foretaste of potential future problems one need look no further than the current
proposals of X3H5 (Parallel Processing), which are that it should develop new Fortran syntax
and semantics (i.e. language extensions) to achieve the appropriate bindings to Fortran 90.
X3 recognised that this would set a precedent and have asked X3J3 for its comments on the
X3H5 proposal (which is for extensions to "ANSI X3.198-1991") before it (or rather SPARC)
makes its decision. Nowhere in the X3H5 proposal, or in the letter from the Chair of
SPARC to Jerry Wagener, is there any recognition of the fact that this might be more than
a domestic U.S. issue. Jerry does refer to any X3J3 response being subject to approval by
WG5 in his paper to X3J3 on the subject, but it is clear that the U.S. Standards hierarchy is
still having difficulty in coming to terms with its international obligations.
However a strong ray of hope has just appeared with a report from Jerry Wagener of the
most recent SPARC meeting, which he attended insofar as it was concerned with the above
X3J3/X3H5 issues. This meeting appears to have been much more positive and
(a) stated firmly that all future bindings should only be to Fortran 90, even if they only
relate to the Fortran 77 elements of Fortran 90 - in other words they are treating the
"archival" status of Fortran 77 seriously;
(b) recommended that the proposed "maintenance release" (or an alternative description
which would require a simpler public review process) should be processed as an "I"
project - which means that a single document is prepared for both US and ISO
purposes (rather than two, identical, documents as with Fortran 90).
I do not have the full details yet, but will try to incorporate them in the version of this report
that is included in the minutes.
Finally it should be noted that Boeing finally withdrew its objection to the ANSI Fortran 90
Standard late last year, or early in 1992, and changed its NO vote to YES (albeit somewhat
reluctantly). This cleared the way for the final processing of the ANSI Standard which is
expected to be approved any time now. A bound copy of the ANSI Standard is to be
presented to Jeanne Adams in Victoria at the WG5 meeting.
Miles Ellis
13 May 1992
Chairman's Report
Good morning. Welcome to this year's annual general meeting. I'm supposed to start by
telling you what an interesting year we've just had.
Well, there were four meetings during the year, starting with last year's AGM. These
meetings have usually been in Mansfield Street, but clearly may not be there in the future.
Apart from the normal business of the group, several topics have been discussed at these
meetings:
Most importantly, we have welcomed the fact of Fortran 90 as a new International
Standard. Soon, we should have the tools available to use Fortran 90, and will then be
able to start coding to the new standard and, I hope, confound all those critics who
claim that Fortran is dead and use C in its place.
We had a short discussion on Data Structures in Fortran 90, demonstrating the language
as being at the leading edge of modern language design using its new abstract data types
and control structures.
We have also discussed a new draft constitution to be presented at this meeting.
Our annual outing this year was to Jodrell Bank, where we saw I how computers are used
in modern astronomy. I found it particularly interesting to see comparisons of tapes from
opposite sides of the world checking the interference fringes of a single photon trying to
decide where to manifest itself.
In general, I would describe it as an interesting and successful year.
Appendix C
BRITISH COMPUTER SOCIETY - FORTRAN SPECIALIST GROUP
Final Accounts 1991 - 1992
GOLD Account
Balance at 30/4/91 1261.30
Interest 117.11
Transfer profit on current and HQ acct + 81.50
-------
to 92/3 year 1459.91
Current Account
Starting Balance 30/4/91 (float) 300.00
Income:
subscriptions (14) 70.00
OUP mailing (last FY) 115.00
Salford mailing 9/91 43.56
(inc 5.65 VAT) ------
228.56 228.56
Expenditure:
To buffet lunch 28 Nov 91 41.97 - .. uncleared
cheque
------
Balance at 30/4/92 528.56
HQ Account
Expenditure: BCS Services (mailings, meetings etc)
Printing 170.66
Post 240.60
Labels 20.72
Temp. help 76.90
------
508.88 508.88
HQ Catering 90.56
VAT on Salford mailing 5.65
------
605.09
Less HQ allocation 91/92 - 500.00
------
Overspend 105.09
Overall profit:
current account less 300 float 228.56
Less uncleared cheque - 41.97
Less overspend HQ account -105.09
------
Balance to Gold account 81.50
Overall Assets:
Gold account 1459.91 + float 300 = 1759.91
Constitution of the Fortran Specialist Group
[This is the Constitution adopted in 1992; it has subsequently been revised.]
1. Constitution
a. The Group is called The British Computer Society Fortran Specialist Group.
b. The Group is formed by a number of people with a common interest in the Fortran language.
c. The Fortran Specialist Group is a part of the British Computer Society.
d. The Fortran Specialist Group consists of the officers listed in Section 3, individual fee paying
members and corporate members.
2. Objectives
a. To promote the use of Fortran.
b. To encourage research into Fortran.
c. To develop competence in Fortran
d. To promote and support the Fortran standard.
e. To provide a forum for Fortran Specialists to talk to one another.
f. To represent the interests of Fortran Specialist Group members with other bodies.
3. Officers
a. The officers are Chairman, Vice Chairman, Secretary and Treasurer, and such other officers as are
required to perform specific tasks.
b. The Chairman of the Group is ex officio a member of the BCS Specialist Groups Management
Committee.
c. The Chairman of the Group must be a Fellow, Member or Associate Member of the BCS.
d. Other officers of the Group should normally be members of the BCS.
e. All officers of the Group must abide by the Code of Conduct relating to members of the BCS.
f. The Group must advise the Chairman of the Specialist Groups Management Committee of
members of the Group who are not members of the BCS.
g. The officers shall be elected by the Annual General Meeting and shall serve from the end of the
meeting at which they are elected until the end of the meeting at which they are replaced.
4. Annual General Meeting
a. Each year the Group shall hold an Annual General Meeting (AGM) in May.
b The Secretary shall send out to all members of the Group, at least 28 days before, notice of the
date, time and place of the AGM, together with the agenda and nomination form for officers.
c. All members have the right to attend the AGM for which there shall be no charge.
d. The quorum is seven members.
e. The agenda for the AGM shall include:
1. The minutes of the previous AGM.
2. The minutes of any EGM held since the previous AGM.
3. The Chairman's Report.
4. The Treasurer's Report and Audited Statement of Accounts.
5. Proposals for alterations to Rules.
6. Proposals for alterations to Fees.
7. Election of Officers and Auditors.
f. The minutes of previous AGM and EGMs must be confirmed, after any amendments, by the
meeting and signed by the Chairman.
g. Any member is entitled to nominate a person for any elected office. Such nominations may be
proposed and seconded at the meeting if not previously received by the Secretary.
h. Every proposal at an AGM shall be decided by a majority of the votes cast by those present
Individual members of the Group each have a single vote. The accredited representative of each
corporate and sustaining member also has a single vote.
5. Extraordinary General Meeting
a. An Extraordinary General Meeting (EGM) shall be convened within five weeks of receipt by the
Secretary of a requisition signed by seven members.
b. An EGM shall transact only such business as is specified in the resolutions or requisitions
convening it.
c. Sections 4b, c, d and g apply also to an EGM.
6. Finance
a. The Group shall follow the BCS Financial Guidelines as issued from time to time.
b. The financial year shall start on 1st May each year.
c. In accordance with BCS Guidelines, the Group shall have at least one Account (Account A) at
Lloyds Bank, Langham Place Branch, used for normal running expenses. For Groups opting out
of the centralised accounting system other accounts may be used for special events or for
investment funds.
d. All cheques drawn on the Group's bank accounts must be signed by two of the Treasurer, the
Chairman, the Vice Chairman or the Secretary.
e. The Chairman is responsible to the BCS for submitting draft budgets, recording ongoing
expenditure and capital expenditure separately for each by 30 November in the preceding year.
f. The Chairman is responsible for making available to the BCS a receipt and payments statement at
the end of every financial year (30th April) in respect of the Group's normal operations and
special events, this statement to be included in the BCS annual accounts subject to audit by the
BCS auditors.
g. The accounts of the Group must be audited each year by an auditor elected at the AGM.
h. All income and property of the Group shall be applied solely to the promotion of the Objects of
the Group.
i. No member of the Group shall receive payment for his services as a member.
7. Relations with the BCS
a. The Group is governed by the rules of the BCS as these apply to Specialist Groups. Where it is
considered that a rule of the Group is in conflict with a BCS rule governing Specialist Group
activities, the BCS rule applies.
b. The Group may use the BCS name to enhance the reputation of their own activities but must not
bring the BCS into disrepute.
c. No member of the Group may speak on behalf of the BCS without proper authority from the BCS.
d. In the event of the winding up or dissolution of the Group any surplus assets remaining after
discharge of liabilities shall automatically rest in the BCS.
Minutes for the AGM business part of the meeting.
Minutes of the Annual General Meeting of the BCS Fortran
Specialist Group held at the Rugby Club of London,
49 Hallam Street, London on 14 May 1992
Present: Tim Bartle Salford Software
Miles Ellis Oxford University
Mike Geary NAG
E Golton RAL
Carol Hewlett LSE
Peter Holland SSL
Chris Lazou University of London
Mike Nunn CCTA
Mike Roth UKAEA
Martin Stanworth Taywood Engineering Ltd
John Young PE-MOD
1 APOLOGIES FOR ABSENCE
Apologies for absence were received from David Muxworthy,
Les Russell and Dave Vallance.
2 MINUTES OF THE LAST AGM
The minutes of the last Annual General Meeting were
presented and approved.
3 MATTERS ARISING
The proposal to spend less time on business and more on
technical matters had not, so far, been implemented.
4 CHAIRMAN'S REPORT
The Chairman, Mike Roth, presented his report which was
reproduced as Appendix B of the full set of minutes already
circulated. The Group held 4 meetings during the year and
published 2 newsletters.
5 TREASURER'S REPORT
The Treasurer, E Golton, presented his draft accounts. He
guided the meeting through the figures which were still
provisional as he was not sure of some of the BCS HQ figures.
He reported that after two years in the job and with the
tightening of financial controls at BCS HQ the finances of the
Group were now under stricter control. If HQ made a transfer,
now he at least, was notified. The final accounts were presented
as Appendix C of the full set of minutes already circulated.
6 ACCEPTANCE OF THE CONSTITUTION FOR THE GROUP
Following the discussion on the Constitution for the Group
it was proposed and accepted unanimously on a formal vote that
the document presented becomes the Constitution for the Fortran
Specialist Group of the British Computer Society.
7 ELECTION OF OFFICERS
The retiring Chairman, Mike Roth, remained in the Chair for
the nomination of the new Committee. The retiring Secretary,
John Young, had received only one nomination for the post of
Chairman; Chris Lazou being proposed by Mike Nunn and seconded
by Les Russell. There were no other nominations so Chris Lazou
was duly elected to the post of Chairman.
For the post of Vice Chairman, John Young was proposed by
Chris Lazou and seconded by Miles Ellis. There were no other
nominations so John Young was duly elected to the post of Vice
Chairman.
There were no nominations for the post of Secretary in spite
of pleas in a Newsletter and on the Nomination Form. As there
was no-one in the meeting prepared to stand for the post of
Secretary it was agreed that the post would have to remain
vacant. The retiring Secretary, John Young, agreed to be Acting
Secretary for the next two meetings.
The Treasurer, E Golton, was prepared to stand again and was
proposed by Mike Nunn and seconded by Chris Lazou. There being
no other nominations, E Golton was duly elected to the post of
Treasurer.
8 BCS BUSINESS
The new Chairman, Chris Lazou, announced that he intended
to seek membership of the British Computer Society which he had
let lapse. Under the new Constitution the Chairman is required
to be a member of the BCS. Miles Ellis and Mike Roth agreed to
sponsor Chris in his application.
The meeting endorsed a vote of thanks to the retiring
Chairman, Mike Roth, for his year's hard work, in particular, in
drawing up the new Constitution. The meeting hoped that he might
be able to attend some of the Group's future meetings. There were
also votes of thanks to the retiring Secretary, John Young and
to the retiring Vice Chairman, Chris Lazou, both of whom remain
on the Committee.
John Young
Acting Secretary
30 July 1992