British Computer Society - Fortran Specialist Group
Minutes of the Annual General Meeting held on Monday April 10th
in the Basement conference Room, BCS Headquarters, 29 Portland
Place, London W.1.
Present:
M R Lewis (Chairman) CRAY Research
D J Holmes Rolls Royce Ltd (Bristol)
J D Murchland University College London
D T Muxworthy Edinburgh R.C.C.
J Roberts-Jones Liverpool City Council
G L Harding (Secretary) European Centre for Medium
Range Weather Forecasts
1. Approval of Minutes of Previous meeting
The minutes of the meeting of 6th February 1978 were approved.
The secretary apologized for the late arrival of the minutes
as this was partly his fault.
2. Matters arising
D T Muxworthy pointed out that he was the Fortran Specialists
Group representative on the BSI programming languages committee
which has formed a working group (BSI DPS 13 WG6) to deliver
FORTRAN comments to the ISO meeting in November.
J D Murchland requested feedback on the form of his abstracts
from the SIGPLAN notices (e.g. level of detail).
It was suggested that the Specialist Group should build up a
library of relevant information such as Manufactures Fortran
reference Manuals, standards etc. Anyone who wishes to donate
such information is invited to contact the secretary.
3. Election of officers for 1978/79
Due to pressure of work and a change of emphasis in his new
position the current secretary resigned at this meeting.
J Roberts-Jones was elected as secretary for 1978/79. His
address is as follows
J Roberts-Jones
Data Processing Division,
City Treasury,
P.O. Box 1,
LIVERPOOL L69 2DQ
Telephone No.: 051-227 3911 Ext. 617
4. Activities of other Fortran Groups
4.1. X3J3
The minutes of the 61st and 62nd meetings
were discussed.
4.2. HECB
The HECB standard Fortran document was discussed.
It was generally thought that these standards are
very restrictive. It was suggested, however, that
such standards do have a role to play. Anyone
who has knowledge of other installation standards
is invited to inform the secretary.
4.3. Canadian Fortran Group
The minutes of their meeting of 1978-02-13 were
discussed.
4.4. ISA
The minutes of their meeting of February 6 and 7
1978 were discussed.
5. Fortran 83
In order to provide a platform for detailed study of the new
standard it was suggested that a working group be set up. This
group could also liaise directly with BSI DPS 13 WG6. Activities
of this group could include submitting proposals for the new
standard to the group for study before sending them to X3J3.
Volunteers to work with this group should contact the secretary.
6. Working Party Reports
There were no working party reports.
7. Other Fortran Events
7.1. Publications
J D Murchland reviewed the February SIGPLAN
notices.
7.2. Implementation Developments
Anyone who is developing a new Fortran compiler
is invited to submit details to the Group.
8. BCS Businesses
There will be a BCS Event entitled "BCS 79 Living with
Computers" to be held January 4th - 6th 1979 at the Institute
for Education. The specialist Groups are organising parallel
sessions. The Group hopes to present one of these sessions and
invites any interested party to submit suggestions for topics
and/or volunteer to speak (for up to 45 minutes).
9. Any other business
It was suggested that the group organises an award for Frank
Engel Jr. This will be brought up for detailed discussion at
the next meeting.
10. Date of next meeting
The next meeting of the group will be on Monday June 5th.
Irene Q Mallgrave of CRAY research will give a talk on the
CRAY-1 Fortran compiler
REFERENCES TO FORTRAN TO RECENT SIGPLAN NOTICES
Abstracted, with some personal selection, by J.D.Murchland, University College
London Transport Studies Group (Gower Street, London WC1E 6BT, 01-387 7050).
SIGPLAN NOTICES 13(1978 February)
6 W.A.Whitaker, the Chairman DoD High Order Language Working Group, replies
to the criticism of the Notices editor in the previous issue that
nothing has appeared in print about the US Department of Defense's
computer programming language design effort, by pointing out that
Communications of the ACM had rejected their article on the subject!
While recognizing that the scale of the DoD project might swamp smaller
language design efforts, they felt driven by economic considerations to
do this, as in other countries (UK, France, West Germany, Sweden, Japan,
and the EEC and CCITT). Comments on the effort are still solicited.
(Contact Lt.Col. Whitaker for details.) The first phase of trial language
designs is to be complete by 1978 February.
8 P.H.Dorn, Dorn Computer Consultants, New York, on the same subject,
remarks that when Ironman was presented at the Share meeting in the
summer of 1977 a good deal of scepticism was expressed: a number of the
design features seemed mutually contradictory, and the target for
execution code efficiency far exceeds what is generally available. Dorn
refers to Datamation's notice (1977 October) that the initial contractors
are SofTech, IntraMetrics, Stanford Research Institute-International
and Honeywell-Bull-CII, all four intending to extend Pascal.
10 D.W.Barron, Computer Studies, University of Southampton: Why go to the
trouble of adding more powerful and useful control and data structures
to Fortran when Pascal already exists?
11 R.J.Cichelli, ANPA Research Institute, Easton, Pennsylvania and Lehigh
University cites the Pascal user's group News, mentions that Pascal on
the Lehigh CDC installation can access Cobol, Fortran and Compass library
modules and asserts that it almost invariably compiles and executes
faster, and takes less space than the other languages.
19 W.A.Whitaker, Defense Advanced Research Projects Agency, Arlington Va:
The US Department of Defense Common High Order Language Effort.
The article gives a history and background of the DoD's language design
effort which has so far produced the 'Revised Ironman' specification.
Much of the DoD's 3G$ a year expenditure on software is for software
embedded in weapon, communication and command and control systems. The
DoD is no longer the dominant purchaser in the hardware market, but it
still has a unique position as a purchaser in the software industry.
High order languages have been little used for embedded systems in the
past because of severe timing and memory considerations. The case for
them is strong, and by the early 1970's each of the military departments
were considering designs. In 1975 a program to set up a single language
was established, one argument being that a single language would have a
full repertoire of supporting software tools, already available at the
beginning of each new project. Since 1976, as an interim measure, only
the languages Cobol, Fortran, Tacpol, CMS-2, SPL/1, Jovial J3 and Jovial J73
may be used, excepted for exceptional circumstances.
The new language specification went through four versions before reaching
'Revised Ironman' (Sigplan Notices 12 (1977 December)). On the way, 23
existing languages were reviewed against the standard (and APL summarily
rejected). Studies showed that if the new language could be produced
it would be suitable, not only for embedded applications, but for all
other user communities (such as avionics, training simulators, scientific
and financial management). Pascal, PL/1 and Algol68 were selected as
base languages for designing the new language, though the four contractors
selected to produce prototypes have each chosen to use the first of these.
Easy generation of efficient compilers for the language on new computers
is of great concern.
It is expected that the language will be available experimentally in
1979 and generally in 1980. Comments and support are solicited. The
publicly available documents on the language effort are cited in Whitaker's
article.
57 G.J.Nutt, University of Colorado at Boulder: A comparison of Pascal and
Fortran as introductory programming languages.
Recently, the Computer Science Department changed from Fortran to Pascal
in its introductory courses, an introductory Pascal course replacing
two different Fortran courses and a Cobol course, to be followed by a
course on larger problems which also introduces Fortran and Cobol. A priori
pro-Pascal views were that it is a more modern language, with better
control structures and encouraging good programming technique, that it
is as easy to teach to beginners as Fortran, and that it is has more
general data types and an extensible data structure feature. Pro-Fortran
views were that (to the contrary) Fortran is easier to teach and learn
than Pascal, that Fortran is far more widely implemented, and has numerous
good textbooks (including Kernighan and Plauger for style) while Pascal
has very few. Also, Pascal is primarily academic and has not been tested
by the real world of applications.
In the comparison, it seemed that Fortran is slightly easier to learn,
mainly because there is less to learn (no extensible data structures,
recursion or arguments passed by name or value). The more numerous Pascal
control structures (While-Do, Repeat--Until or For -Do) were no harder
to grasp, and Pascal input-output was much easier for purely numeric
data than Fortran Formats. Nutt felt that the Pascal programs were better
structured, easier to read and took less time to write.
It was concluded from the experiment that Pascal should be taught to
students first, and then those who needed it should go on to a short
Fortran course. Problems with Pascal at this time are the lack of Pascal
textbooks, and character set difficulties (on the CDC 6400 system).