Minutes of the BCS Fortran Specialist Group Meeting
held in R68 building CR13 at the Rutherford Appleton Laboratory,
Chilton, near Didcot on 25 October 1990
Present: Miles Ellis Oxford University
Ikko Evans Queen Mary & Westfield College
E Golton RAL
Peter Holland SSL
Mike Nunn CCTA
M J Roth AEA Technology
John Stratford Queen Mary & Westfield College
Paul White Met Office
John Wilson Leicester University
John Young PE-MOD
1. APOLOGIES FOR ABSENCE
Apologies for absence were received from Mike Geary, David
Holmes, Chris Lazou, David Muxworthy, John Reid, Les Russell and
Lawrie Schonfelder.
2. MINUTES OF THE LAST TWO MEETINGS
The name of John Bruce should be added to the list of
apologies for absence for the meeting of 10 May 1990.
David Muxworthy has written to the Secretary, John Young,
on the following point:-
The report on the WG5 meeting of section 4 of the minutes
of 10 May 1990 lacks the important point that the two countries
that voted "no" agreed to change their vote to "yes" in the light
of changes to the document agreed at the meeting. I would
propose deleting the sentence "This is a sufficient .....
Standard" and adding instead "After changes agreed at the WG5
meeting these two countries agreed to change their votes to YES".
The meeting accepted this amendment to the minutes.
In the minutes of the meeting of 28 June 1990 in section 4
the sentence "The DISC processing of the Standard was just
starting." should read "The DIS processing of the Standard was
just starting".
There being no other changes and following a decision taken
at the 10 May 1990 meeting the Chairman, John Wilson, signed
copies of the Minutes of the Meetings of 10 May and 28 June 1990.
3. MATTERS ARISING
Scottish Sub-Group
The Secretary had received a letter from the Sub-Group
Convenor, John Bruce, resigning from the post. John has offered
to continue until a successor is found. He also reported that
the meeting arranged at the Dundee College of Technology had
failed altogether. This was partly due to problems within the
local Dundee Branch of the BCS.
The Chairman had received an e-mail from David Muxworthy
reporting on the Glasgow meeting held on 22 October 1990. Ian
Cheyne of Scottish Power and Jamie Reid of NEL gave two excellent
talks but regrettably only 9 people attended. The next meeting
for January/February 1991 will be about Fortran 90 compilers and
tools.
Dynamic Strings in ISO Fortran
E Golton requested a copy of this technical paper by Lawrie
Schonfelder and J S Morgan. The Chairman said that the paper had
been reviewed and refereed but that a copy was not yet available.
4. REPORT FROM X3J3 REPRESENTATIVE
The X3J3 representative, Miles Ellis, gave a general report
on the progress of the Fortran 90 Standard since May. There had
been a number of meetings during the summer but the two important
ones were the WG5 meeting in Rotterdam and the X3J3 meeting in
Oxford during August.
As reported in the last minutes the May meeting of X3J3 had
passed the Standard to ISO and the DIS ballot was supposed to
start on 1 July 1990. For various reasons the DIS was not sent
out by ISO until 2 August 1990. Hence the DIS ballot ends on the
2 February 1991.
The WG5 meeting in Rotterdam was mainly concerned with
editorial processing. However, the meeting spent a long time on
the Forward Reference Issue. This issue concerns whether a call
to an internal procedure within a procedure in Fortran should be
declared "up-front" making things easier for compiler writers.
There was disagreement over the extra compiling time required if
a statement like FORWARD was not included in the Standard varying
from 2% to 50%. In particular, manufacturers did not feel any
increase would make much difference. Straw votes in the
Rotterdam meeting were strongly against the inclusion of Forward
Referencing and it was clear that WG5 did not want this major ~
item. This change to the Standard would have to be done now or
not at all.
For a second time WG5 operated as a technical committee with
its major work being editorial changes.
The following week X3J3 met in Oxford. Again the Forward
Reference Issue was raised but again was not passed due to
insufficient support. X3J3 reviewed the list of editorial
changes made by WG5 and slightly changed it. The majority of the
time was taken up with the processing of letters from the second
public review and dealing with the complaints that the first
public review comments were not published by the time of the
second public review. Miles had formally proposed that just one
single letter be sent to people but was ruled out of order.
Following the X3 decision to keep Fortran 77 as a separate
standard and to have Fortran 90 as an extended Fortran Standard
it was now theoretically possible with the publication of the
international Fortran 90 Standard for there to be three US
standards. In fact, the wording of a formal announcement of the
new Standard by X3 as a news release guarantees that this
situation will arise.
The US first (or third?) public review period of Fortran 90
ends on 19 December 1990 with the near certainty that the
National and International Standards will be different. X3J3 are
very upset by this situation. However, the US is not likely to
have a different Fortran 90 Standard to the rest of the world and
it is X3's intention for it to be the same. SC22 had requested
that all countries withdraw Fortran 77 as the Standard but the
US will make Fortran 77 a US Archival Standard.
In summarising, Miles stressed the following points:-
... The Draft International Fortran 90 Standard is half-way
through its six month ballot.
... The WG5 meeting in March 1991 will review comments from the
ballot.
... All countries (except the US) should vote YES.
... Any NO vote would be on editorial changes only.
... The X3J3 meeting in April 1991 will complete the final
Fortran 90 Standard document.
... The International Standard is due to be published in Summer
1991.
... The next X3J3 meeting is to be held during the week before
Christmas 1990 and could be faced with a possibly
significant list of changes from the US public review.
... The UK may vote NO but did agree with WG5 list.
... The BSI ballot was sent out without any changes.
... It was proposed that the Group should endorse the BSI
action.
... The Fortran 90 Standard would be adopted as the British
Standard in November.
... The BSI position is that international standards should be
developed internationally.
... WG5 is to establish its position as the next Fortran
Standards Committee.
... X3J3 will have completed its task but could easily be re-
convened if necessary.
Any comments on the Fortran 90 standard should be sent to
the Chairman of the BSI Fortran Panel, David Muxworthy.
The BSI initiative to transfer programming languages to DISC
has failed to "get off the ground". The BSI is seriously
considering that programming languages may be moved to another
body, for example, the BCS Fortran Specialist Group. The meeting
felt that if this situation did arise the Group would be in a
good position to cope with the extra responsibility.
5. 21st BIRTHDAY CELEBRATIONS
The convenor for the 21st Birthday Celebrations, Mike Nunn,
told the meeting that the decision to hold a Day meeting at
Imperial College was not possible as the room was only available
in the evenings. The brochures concerning the room availability
and dinner arrangements should have been circulated at the last
meeting. There followed a discussion on why the Group was
organising this meeting and who was likely to attend. The
meeting could not remember the rationale for proposing such a
celebration particularly as the Group had reviewed its first
twenty years with an excellent talk from David Muxworthy. It was
agreed that February 1991 was too close to organise the
celebrations and that it would have to be an extra meeting
anyway. It was far too late to book a key-note speaker even if
the Group had thought of somebody to invite.
The meeting thought that 25th Birthday Celebrations would
be much more appropriate. On a straw vote 2 voted for, 4
against, and 4 abstained for continuing with the 21st Birthday
Celebrations.
6. DATES OF FORTHCOMING MEETINGS
The meeting endorsed the quarterly pattern of meetings
proposed for 1991 with the exception of the August meeting which
is moved to early September. It was decided to continue to meet
on a Thursday. With the move of BCS HQ still very uncertain the
venue of the November meeting was left undecided.
The following meetings are planned for 1991:-
Thursday, 7 February, 1991. BCS HQ.
Compilers for Fortran (Tentative Title).
Thursday, 16 May, 1991. BCS HQ.
AGM at 2 pm followed by Report and Debate on Fortran 90.
Thursday, 5 September, 1991.
Proposed visit to Jodrell Bank.
Thursday, 28 November 1991.
Venue to be decided.
7. ANY OTHER BUSINESS
Talk on FORCHECK by Erik Kruyt.
The Chairman had received a letter from John Appleyard of
Polyhedron Software suggesting that the Group might like a talk
from Erik Kruyt on Forcheck. The meeting decided that as the
Group had several talks on similar packages in the recent past
that an extra meeting was not justified. It was suggested that
any member who was interested should attend the talk to be held
at Rutherford Appleton Laboratory on 6 December 1991.
UNICOM Seminars Limited
The Chairman, John Wilson, had received information from
Unicom who organise seminars of Information Technology topics and
enclosing details of various forthcoming events. Unicom invited
the Group to collaborate on a seminar to be called "The Visual
Computer: Designing, Visualising and Animating with Computer
Graphics". Unicom also requested an opportunity to distribute
leaflets to members.
Fortran 90 by Michael Metcalf and John Reid.
The Secretary, John Young, had received a request from
Oxford University Press to include a one page flyer in the
Group's next mailing for the book Fortran 90. It was pointed out
that this is not a new book but an update of Metcalf and Reid's
book "Fortran 8x explained". The meeting approved the flyer at
the usual cost of mailing to members.
Achievements with Software Engineering
The Chairman, John Wilson, had received notice about the
above free one day seminar to be held on 22 November 1990 and
organised by Salford University Business Services Limited and the
DTI.
Signatories to the Group's Bank Accounts
The meeting resolved that the names of the authorised
signatories to its bank accounts should be changed to the names
of the current Chairman, Secretary and Treasurer.
Specialist Groups' Management Committee
John Wilson attended the recent Committee meeting and
summarised with the following points:-
(1) There was no news of the move of HQ but BCS were still
trying to sell the lease.
(2) There was a lot of criticism of HQ facilities. Chris
Maskens had been appointed temporary liaison office for
Specialist Groups. He would be dealing with the Group's
cheques and the Membership List.
(3) A new Specialist Group for POPLOG and POP Languages called
PLUG had been formed. A Group for Computers and Writing
was being planned.
(4) BCS is a founder member of CPEIS and is now attending
meetings.
8. VISIT TQ THE SPACE DATA CENTRE AND THE ATLAS BUILDING FOR
TALKS AND DEMONSTRATIONS BY RAL STAFF
The morning session concluded with a short tour of the Space
Data Centre hosted by the Treasurer, E Golton.
The afternoon session consisted of three talks by RAL staff
followed by a tour of the Atlas Building, in particular, the
computing facilities. Dr T Daniels gave the introductory talk
followed by Dr J Gordon who gave a talk which included RAL's
approach to Parallel Processing. The final talk on Video Output
was by Dr C Osland who also showed the Atlas Video Facility
Demonstration Video tape. Details of the Atlas Vide Facility are
included as Appendix A.
John Young, Secretary
12 December 1990
ATLAS VIDEO FACILITY
The Atlas Video Facility allows SERC users - and those of other Research Councils - to gen-
erate animation sequences on normal videotape. As such, it replaces the film output facility
that used to be provided at the Atlas Centre. However, the technology by which the video
system is implemented means that many more facilities can be provided directly, such as
several sequences at different speeds - all from one original sequence.
HARDWARE
The video system is now based on an Abekas A60 digital video disk - a standard broadcast
quality machine - which stores 30 seconds (750 frames) of video. Pictures are put onto the
disk by a Primagraphics computer which generates them from files sent to it by users. The
Primagraphics system contains a full broadcast quality, 24-bit framestore, genlocked to a
studio quality sync source; this allows all equipment to be interconnected without problems.
The Abekas may be used to record frames (25 per second in the UK) or fields (50 per second
in the UK).
The Abekas disk allows repeated sequences to be produced at normal speed, faster or slower
than normal, or even reversed. For instance, a cyclic sequence of pictures only needs to be
stored once and can be recorded onto videotape many times for convenient viewing.
The output from the Abekas disk is finally written to a Sony U-Matic Hi-Band SP recorder
under the control of a Lyon-lamb animation controller, which allows many different
sequences to be combined, with pinpoint accuracy, onto a single tape.
Users may take the Hi-Band U-Matic SP master tapes or master tapes made on Lo-Band
U-Matic or VHS formats if they prefer. Direct output from the Abekas disk to other formats
(one inch, Betacam, D2) is possible if the relevant recorder can be brought to RAL.
SOFTWARE
All the software for the video system has been developed by the Atlas Centre. Most users
send files to the Atlas Video Facility in ISO Computer Graphics Metafile (CGM) format.
Alternatively, pictures may be sent in raster form; this is appropriate when producing com-
plex pictures with shading and ray-traced images.
The video system software controls the production of pictures in the Primagraphics com-
puters framestore and the transfer of these to the Abekas disk. When a sequence is complete
in the Abekas, the software then simultaneously instructs the Abekas to play it back and the
Lyon-Lamb animation controller to record it onto videotape.
POST-PRODUCTION
In many cases, users require not just a single sequence, but a set of sequences. together with
captions, titles and possibly soundtrack and musical background and/or effects. Where the
demands are quite simple - for instance, some title frames and a spoken commentary, this can
be done within the Atlas Centre. For more complex editing - such as some of the effects seen
on the sample videotape - we are able to use the Hi-Band editing suite in RAl_'s
Reprographics section.
ATLAS VIDEO FACILITY
DEMONSTRATION VIDEOTAPE
This tape shows a number of sequences CRANKSHAFT LUBRICATION
produced by users of the Atlas Video Facility
over the last year. All sequences were prod- This pilot sequence was_produced by Dr. Jack
uced on the facility apart from the opening Fendley and others at King's College, London.
sequence - recorded off air from a BBC-tv It shows the pressure profile in the oil of a
transmission - and the short sequence from crankshaft bearing. The output was produced
"The Conquest of Form" by William Latham. in CGM format on a VAX computer and sent
Most Sequences were produced by Computer to Atlas on a magnetic tape. This video was
Graphics Metafiles (CGMs) and some from also shown at Kings College Open Days
raster files; one (the Indian Ocean Triple earlier this year.
Junction) was produced directly by the video
system computer itself. INDIAN OCEAN FLOOR
This sequence shows a simulated flight
FRAM - oceanographic modelling around the ocean floor at a junction of three
tectonic plates in the Indian Ocean. The data
The first two sequences were produced by for the pictures was collected by GLORIA, a
programs developed at the Atlas Centre from submersible sidescan sonar system operated
code provided by the FRAM (Fine Resolution by the Institute of Oceanographic Sciences
Antarctic Model) Project, organized by Deacon Laboratory. The data was collated
NERC. They show a view of the whole area by Tim Le Bas of IOS and the video was
covered by the model and a close-up of the made directly by a program running in the
Agulhas Current region. CGMS were prod- video system computer. This sequence was
uced on the Silicon Graphics IRIS. The final first seen at the international Oceanology
two sequences show the equivalence between Conference in March and has been seen at
a polar and a Mercator projection and a tran- many conferences and exhibitions since.
script from a session on the Silicon Graphics
IRIS; all are used by the Open University in UGAMP - atmospheric modelling
their new Oceanography course. This sequence describes a number of simu-
lations of how CFCs might disperse in a gen-
FLUID DYNAMICS eral atmospheric model. The programming
was done Simon Cooper of the UGAMP
These two sequences show the flow of water (UK Universities Global Atmospheric
into a semi-closed container and the inter- Modelling Project) and the main code was run
action between a wave and static water on a on Cray supercomputers at London and Atlas.
seashore. The program was written by Siebe The resulting data was processed to produce
Jorna of St. Andrews University and run on CGMs on a Sun workstation at Atlas.
the Silicon Graphics IRIS at the Atlas Centre
to produce CGMs. He wanted to produce WILLIAM LATHAM
sequences that could be used in a Fluid
Dynamicsccourse and so the models are sim- William Latham's sculptures are designed by
ple and the animation slow. him on the IBM 3090 at the Atlas Centre,
where he does much of his work. The final,
TABLE LAYOUT fully rendered frames are then produced at
Atlas and put onto video tape here or at the
This very simple sequence was produced by IBM Scientific Centre at Winchester.
Prof F R A Hopgood on a Sun workstation William's second film "The Evolution of
using SUNPHIGS. Qutput from the Sun, in Form" won first prize for technical innovation
CGM format, was sent straight to the video at this year's IMAGINA competition in
system where it was made into a video tape Monte Carlo. An exhibition of his films and
in about four minutes. electronic sculptures is currently at the
Natural History Museum in London.