BRITISH COMPUTER SOCIETY FORTRAN SPECIALIST GROUP
Minutes of a Meeting held on
Friday, 27th July, 1973 at
Barnard's Inn, Holborn Bars,
London EC1 at l0.30 a.m.
Present: Mr B.H.Shearing (Chairman) Alcock Shearing & Partners
Mr P.D. Bond Philips Industries
Mr P.A. Clarke Rothamsted Experimental Station
Mr J.C. Cullen B.P.
Dr A.C. Day University College London
Mr D. Ellison University of Manchester R.C.C.
Mr P.J. Hammond B.P.
Mr I.D. Hill M.R.C.
Dr C.B.A.Price Honeywell
Mr K. St.Pier G.E.C.
Ms. S.M. Sinclair Atlas Computer Laboratory
Ms. M. Spooner Computing
Mr T.L.van Raalte A.W.R.E. Aldermaston
Mr D.Winstanley University of Birmingham
Mr D.T. Muxworthy (Secretary) Edinburgh Regional Computing Centre
Apologies for Mr A. Bruce Government Communications H.Q.
Absence: Mr T.D. Palmer N.C.B.
Professor J.S. Rohl University of Manchester I.S.T.
ACTION
1. APPROVAL OF The minutes of the meeting of the 18th May
MINUTES 1973 were approved, subject to the
substitution of 'Hammersley' for
'Hemmereley' in paragraph 2b.
2. MATTERS ARISING a. The Chairman and Secretary were to BHS/DTM
FROM THE inform the Editor of the Computer Journal
MINUTES of the latest estimate of the publication
date for the draft proposed American
National Standard (dpANS) for Fortran
(see section 5 below).
b. It appeared that there was now no medium
for publishing the Group's annual report
which formerly appeared in the Computer
Bulletin.
3. USE OF The Chairman had been informed that the
SPECIALIST advertising plan was introduced to
GROUPS FOR forefend other, less controlled, schemes
ADVERTISING and, having been reassured that information
was not distributed outside the BCS, that
members could withdraw their names, that
only conferences, books and the like would
be advertised and these only after the Group
Chairman's approval in each case, he (Mr
Shearing) would now inform Mrs Roberts that BHS
she may send relevant material to him for
consideration for distribution.
4. ECMA FORTRAN Dr Day had written to ECMA (minutes of 18.5.73
STANDARDS section 8) and M. Besse, the ECMA Deputy
ACTIVITIES Secretary General, had responded by offering
to exchange relevant documents. The
Secretary was to arrange this. DTM
ECMA committee TC8, which produced the
ECMA Fortran standard of 1965, reopened
its activities in February 1973 with
representatives from six European based
computer manufacturers, plus Honeywell
and Univac; the chairman is Mr van
der Munnik of Philips-Electrologica. The
committee had established contact with
X3J3 and its program of work was to
review the X3J3 proposals for consistency,
relation to common practice, relation to
X3.9 and implications for users;
consideration of relation to ECMA-9 and
X3.l0 was deferred.
At the second meeting in April a number
of X3J3 proposals were selected for
further discussion; several items already
approved by X3J3 were opposed by TC8.
5. AMERICAN The minutes of the X3J3 meeting on May 15-18
FORTRAN 1973 were considered. The principal
STANDARDS decisions then taken were:
ACTIVITIES
a. To rescind apostrophes to delimit
Hollerith constants, thereby distinguishing
syntactically Hollerith and CHARACTER
constants.
b. To rescind the PUNCH statement.
c. To allow CHARACTER functions.
d. To include new intrinsic functions and
generic function names for the nearest
integer function.
e. To make several clarifications to the
direct access input-output proposals.
Subject to checking by the secretary, it was
decided to send comments on current proposals
to X3J3; further details are given in Appendix
1 to these minutes. Subsetting of the new
standard is described in section 6 below.
The estimated date for the publication of
the draft proposed standard was now the first
quarter of 1974. Mr Noll, the X3J3 member
with special responsibility for international
liaison, had recently indicated that comments
from this group were still most welcome,
especially when they pointed out irregularities
or inconsistencies, but it was now probably too
late to argue points which were mainly a matter
of taste. Contrary to information available
at the previous meeting, Mr Noll was optimistic
about the draft being approved at an early date.
6. SUBSETTING The current thoughts of X3J3 on subsetting
were that there should be three levels,
viz Fortran, General Fortran and Fundamental
Fortran the criteria for inclusion of
elements into which are shown in Appendix
2 to these minutes. Mr Winstanley presented
a criticism of the proposals; his main
points were that:
a. Fundamental Fortran was still a large
language, significantly bigger than ISO
intermediate level.
b. That the two upper levels were too
close, there being only nine elements of
the language in Fortran but not in General
Fortran and that therefore General Fortran
should be revised to be closer to X3.9.
c. That direct-access input-output should
not be required at lower levels.
d. That significant savings of space could be
achieved by including list-directed and not
formatted input-output at the lowest level.
It was decided to send comments to X3J3. DW/BHS/
Mr Winstanley was thanked for his work. DTM
7. OTHER a. Attention was drawn to a BCS conference
BUSINESS on Job Control to be held in January 1974.
b. A new specialist group on pterotechnology
was being formed; details were available
from BCS HQ.
c. A room at BCS HQ was again available for
holding meetings. A majority at the meeting
favoured using that rather than Barnard's
Inn.
d. Talks given to Study Group 5 during the
year were to be published in the early
autumn.
e. Attention was drawn to a paper entitled
"Serious Fortran" by J. Larmouth, the first
part of which was published in Software
Practice and Experience Vol 3 pp. 87 - 107
and the second was to appear in the following
issue.
f. It was noted that a manual on BASIC had
been published in the NCC Standards series.
Anyone interested in BASIC was asked to
make themselves known to the chairman.
8. DATE OF NEXT The next meeting will be held at 10.30 a.m. on
MEETING Friday, November 9th, 1973 at BCS Headquarters,
29 Portland Place, London W.l. (Please note
the change of venue).
COMMENTS ON X3J3 PROPOSALS
l. Concern was expressed that rescinding apostrophes to delimit
Hollerith strings would remove this facility also from
format field descriptors. In fact the facility remains in
the guise of a CHARACTER field descriptor.
2. It was decided to ask whether logical unit zero had special
significance. It does not.
3. It was decided to ask for clarification of the effect of
e.g. (T50/...). This is now specifically allowed to guarantee
that the first record output has length at least 49. The
actual length of a formatted record "depends primarily upon
the number of characters put into the record when it was
written. However, it may be dependent upon the processor
and the external medium".
Thus (8Fl0.5, T201) used with a punched card would cause an
error condition, just as (20Fl0.5) would.
4. It was decided to suggest that UNIT=unit be allowed also in REWIND
etc. However, unit is now proposed to be a positional parameter
in all input-output statements except INQUIRE (where UNIT= is
optional).
The only remaining comment, which in itself was not worthy of a
memorandum to X3J3 was:
5. The Group still disliked a variable apparently being assigned
on the right of an equals sign, as in OPEN=L.
Criteria for Inclusion of Elements of FORTRAN into
Various Subsets
FUNDAMENTAL FORTRAN
l. That this subset be proper subset of GENERAL FORTRAN.
2. That this level correspond approximately to ISO RECOMMENDATION
R1539 Intermediate Level FORTRAN.
3. That this level include (at a fundamental level) those features
of the full FORTRAN language which significantly increase the
scope of the language.
4. That the elements of this subset make a minimum demand on primary
storage space requirements particularly at run time.
5. That this subset require a minimum of effort for the development
and maintenance of viable FORTRAN processor.
GENERAL FORTRAN
1. That this subset be a proper subset of FORTRAN.
2. That this level correspond approximately to the ANS X3.9-1966
FORTRAN language.
3. This level should have the X3.9-1966 standard as a proper subset
and incorporate those features of the new standard which regularize
anomalies which previously existed.
4. That this level include (generally) those features of the FORTRAN
language which will significantly increase the scope of the
language.
FORTRAN
l. That this level represents the full standard as defined by X3J3.