BRITISH COMPUTER SOCIETY FORTRAN SPECIALIST GROUP



Minutes of a meeting on

Wednesday 19th April 1972

at B.C.S. Headquarters,

29 Portland Place, London

W1 at 10.30 a.m.



Present:   Mr J.S. Gatehouse (Chairman)    G.E.C.

Mr J.C. Cullen                  B.P.

Dr A.C. Day                     University College, London

Mr P.J. Hammond                 B.P.

Mr I.D. Hill                    M.R.C.

Mr I.D.K. Kelly                 A.I.A. Computing Ltd.

Mr D.J. Maisey                  I.C.L.

Mr R. Mankin                    Olympic Computer Services

Dr C.B.A. Price                 Honeywell Information Systems

Mr K. St Pier                   G.E.C.

Mr B.H.Shearing                 Alcock Shearing & Partners

Mr D. T. Muxworthy (Secretary)  Edinburgh R.C.C.


Apologies for   Mr E. Bodger               IBM Information Services Ltd.,

Absence:        Mr P.D. Bond               Philips Industries

                Mr R.E. Day                Edinburgh R.C.C.

                Mr D.H. Marwick            Heriot-Watt University

                Mr K. Normington            Lanchester Polytechnic

                Prof. P.A. Samet           University College, London


            Action


1.      APPROVAL OF        The minutes of the meeting held on Thursday

        MINUTES            16th December 1971 were approved.


2.      MATTERS ARISING    Mr Gatehouse had read a draft of the

        FROM THE MINUTES   book "Computer Programming Made Simple":

he said that it was concerned largely

with teaching programming from a terminal

and he had asked the publishers to make

clear that the Fortran used was non-standard.


3.      OFFICERS FOR       The present officers indicated that they

        THE COMING YEAR    were willing to serve for another year and

were duly reappointed. Mr Gatehouse

did not wish to continue as chairman after

the 1972-73 session. It was decided to

create a new office of vice-chairman and

Mr Shearing agreed to serve in this position

for a period of one year and on the

understanding that the vice-chairman was not

expected to succeed to the chairmanship.


The officers for 1972-73 are thus:


Chairman:                Mr J.S. Gatehouse

Vice-chairman:           Mr B.H. Shearing

Minutes & meetings

Secretary:               Mr D.T. Muxworthy

Document

Secretary:               Mr R.E. Day


It was decided to request the editor of               DTM

the Computer Bulletin to print both the

Chairman's and Secretary's names and

addresses in the Group's entry in the

list of specialist groups.


4.      BUDGET FOR         The Chairman read a memorandum on the

        THE COMING YEAR    reorganization of BCS staff functions;

there had been no other indication of a

reduction of income and he had therefore

prepared a budget which allowed for use

of a meeting room on five half-days per

year and occasional duplicating and

mailing services. The estimated income

was nil, estimated expenditure £50.

(In 1971-72 the estimated expenditure

was £50, actual expenditure £10. The

budget does not cover normal mailing of

minutes and agenda which is financed by

Edinburgh R.C.C.). At the same time the

BCS had asked about the types of meetings

the Group could organize. The reply had

been that a variety of meetings could be

held but there were no plans to deviate

from the present pattern.


5.      ANNUAL             A brief contribution to the Society's

        REPORT             annual report was required by May 2nd.                DTM/

The secretary was asked to draft this                 JSG/BHS

and submit it to the chairman and

vice-chairman. The year was taken to

be May 1971 to April 1972 and the report

was to mention (a) contacts with ANSI,

(b) that the extensions working party

was acting as a successful ginger group,

(c) that the group's reports had been

well received by ANSI and (d) that

interests had broadened during the year.

The edited report appears in Appendix A

to these minutes.


6.      PROGRESS ON        The Secretary reported that he was now

        THE ANSI STANDARD  on the ANSI X3J3 active members mailing

list and summarized the latest information

on progress with the standard. Fuller

details are given in Appendix B to these

minutes. The ANSI procedure is that

subgroups present their conclusions to the

whole X3J3 Subcommittee for voting; approval

at this stage implies intention to include

an item in the standard but cannot imply

certainty because of possible ambiguities

or contradictions found subsequently.


7.      REPORTS FROM THE   a. Extensions Working Party. Mr Shearing

        WORKING PARTIES    said that a report had been sent to ANSI in

November which contained some detailed

proposals and attached the working party's

priorities to each item of the ANSI

extensions list. It had been decided that

the most effective way of proceeding

was to hold meetings about 4 weeks

before each X3J3 meeting and to comment

on the latest proposals, particularly

to point out areas of strong disagreement.

One such meeting had been held. Concern

was felt especially about regularity of

the new language. It was anticipated

that the party would have a considerable

amount of work in vetting the draft standard

when it was published.


This meant that the intention to publish

a consensus report in the Computer Bulletin

had been temporarily abandoned. However,

in order to stimulate the flow of Fortran            (EB/ACD)

papers, authors of the detailed sections

of the November report were encouraged to

submit them for publication as personal

contributions. Any surplus reprints of

Fortran papers could be sent to the

secretary for forwarding to ANSI etc.

Dr Price asked that the ANSI papers be

circulated more widely in the Group but

it was felt that this was not practicable.

Members not in working parties could help

by making themselves available for

consultation.


b. Free Format Working Party. Mr Marwick

reported in writing as follows. 'We have

received a number of comments as a result

of the circulated letter and the letter

in the Computer Bulletin. While the

response has not been outstanding, there

is probably enough material to produce a

report. We hope to do this in the near

future and a copy will be sent to all

members of the Group and everyone who

expressed interest in the project'.


It was suggested that the report be                   DHM

submitted for publication, either as

a personal contribution or with the

support of the Group after members had

had opportunity to comment on it.

Mr Marwick had also corresponded with

BSI about standardizing free-format

input-output. He had been invited to

submit proposals for consideration but

had been informed that the only likely

course of action by BSI was to forward

them to the appropriate ISO Secretariat

(ie ANSI).


c. Small Machines Working Party. No

report was available; this working party

was believed to be dormant.


d. Diagnostics Working Party. Mr Hill

reported that Professor Samet had resigned

the chairmanship of this working party

and that he (Mr Hill) was unable to take

it up although both wished to remain

members if a chairman could be found.

No one at the meeting expressed interest.


It was therefore decided to ask (via these            All

minutes) for a volunteer to chair this

working party.


8.      SOFTWARE 72.       Mr Gatehouse had been invited to speak

on Fortran at the Software 72 conference

in Canterbury in July. He planned to

talk mainly about usage in the U.K. and

not to dwell on proposed extensions. Any              All

points which members of the Group would

like to be covered should be sent to

Mr Gatehouse.


9.      HIGH LEVEL         Mr R.E. Day has represented the Group

        LANGUAGE           on the organizing committee for the

        CONFERENCE         conference 'High level languages -

the way ahead' and he provided the

following information. The conference

will be held at York University on

October 4 to 6 1972 starting at 2.30 p.m.

on the first day. The principal

existing languages, and a few minor ones,

will be surveyed and compared and

requirements for a future language will

be discussed. A number of speakers of

international repute have been invited

to take part. There will be a dinner

on the first evening: the al1-inclusive

cost for BCS members will be £28. Further

details are scheduled to be published in

the June 1972 Computer Bulletin. The

speaker on Fortran will be Mr B.H. Shearing.


10.     COMMUNICATION      Besides Mr Marwick's communication with

        WITH BSI           BSI (section 7b above) the Secretary had

requested the BCS Standards Committee to

forward to BSI the Group's recommendation

that the ISO Fortran Recommendation be

reissued as a British Standard; no reply

had been received.


11.     FUTURE             Due to other commitments Dr Day had been

        ACTIVITIES         unable to arrange the lectures mentioned

at the previous meeting and there was some

discussion as to whether the Group's desire

to be outward-looking would be better

fulfilled by lectures or by more publications.        ACD/

Dr Day offered to act as clearing-house for           All

ideas on this subject. (His address is

University College Computing Centre,

19 Gordon Street, London WC1H 0AH). The

two original talks have been postponed

indefinitely. It was suggested that

the Group could give formal support to

anyone giving a lecture on an appropriate

subject.


12.     OTHER              a. The Chairman read a letter to the BCS

        BUSINESS           Secretary - General from Mr J.F. McClary

of Los Alamos Scientific Laboratory which

praised the discussion of Fortran in the

Computer Bulletin and expressed a desire

to establish links between the Los Alamos

Fortran Committee and the Group. The

Chairman was to reply to this letter. (In             JSG

the reply collaboration with Los Alamos

was welcomed; it was also pointed out

that the Group was in close touch with

ANSI X3J3).


b. The Chairman read a letter asking for

volunteers to examine file-handling projects

in Fortran and/or BASIC for the City and

Guilds Basic Certificate in Computer

Programming. Two volunteers were found

at the meeting and it was suggested that

staff at the University College Computer

Centre might be able to help. The Chairman            JSG

was to reply to City & Guilds. The letter

is reproduced in Appendix C. Any member of

the Group who is interested should contact

Miss Hoban directly. The fee is understood

to be £2 per candidate plus travelling

expenses.


13.     DATE OF            The dates for the next two meetings

        NEXT MEETINGS      are Thursday July 6 1972 and Thursday

September 21 1972. The July meeting

will be cancelled if there is

insufficient business. Both meetings

will be held in the Council Room at

B.C.S. Headquarters. (Secretary's

note: please put these dates in your

diary NOW).




APPENDIX A


CONTRIBUTION TO THE BCS ANNUAL REPORT


        The Group has held three meetings during the year and its working

parties have also been active. One of the Group's principal objectives,

the influencing of the forthcoming new American National Standard for

Fortran, has been pursued and there is now a full flow of information

between the ANSI Fortran subcommittee and the Group. The Group's

submissions have been well received in the U.S.


        The Group has also broadened its interest during the year and is

becoming a focus of interest and an independent source of information on

Fortran matters. This role is expected to increase as it becomes

generally realized that the new standard is to be a major extension of all

current versions of the language. The Group has been encouraged by the

recent surge of articles, letters and conference papers which clearly show

a continuing growth of interest in Fortran.




APPENDIX B



Subject: Status of Proposals to Extend and          February 18, 1972

         Revise ANSI Standard Fortran


From: J. C. Noll


To: SHARE Fortran Project



        The American National Standards Committee X3J3 is currently

revising and extending the current ANSI Fortran Standard. The

attached list of proposals for revising and extending standard

Fortran is taken from a working document prepared for use by ANSC

X3J3. The titles of some proposals are somewhat cryptic, since

they are short working titles used and understood by X3J3

members.


        The proposals are separated by group assignment and sorted

by priority rating within each group. The meaning of the

status/priority is:


Approved: Approved by formal vote of full committee.

Not Recommended: Not recommended by the group.

Disapproved: Disapproved by vote of full committee.

Other status: Status or priority assigned by the group.


        Please note that the status of these proposals is changing

rapidly. Furthermore, any proposal approved (or disapproved) by

the full committee can be disapproved (or approved) by subsequent

action.


        If there are any questions or comments on any proposal

please contact me or any other member of ANSC X3J3. Feedback is

welcomed.



HO-8241-JCN                        J. C. Noll

Manager, SHARE Fortran Project

Voting member ANSC X3J3


Att: As above



EXTENSION AND REVISION PROPOSALS FOR ANSI STANDARD FORTRAN

February 18, 1972



GROUP 1 - PROGRAM FORM AND PROGRAMS, FORMAT STATEMENT


Approved


        16         Program statement to identify main programs.

        57         Additional special characters.

        86         Allow more parenthesis levels in formats.

        161        Allow T format descriptor.


Should be in Standard


        25         Ignore trailing blanks in input data.

        84         Ignore embedded blanks in numeric input data.

        100        Specify action if output number exceeds field width.

        109        Provide numeric output for std. info. interchange.

        133        Specification of integers in formats during execution.

        144        Do not allow non-Fortran procedures in std programs,


Recommended


        5         Specify a statement delimiter character. allow more than

                  one statement per line.

        43        Explicit format descriptor to specify rescan point.

        59        Embedded comments,

        95        Allow G format descriptor with any arith. variable.

        98        R Format fields descriptor,

        117       Allow non-Fortran chars. on comment lines.


Of Some Value


        118       Make end line a statement.


Of Trivial Value


        96        X Format descriptor skips chars. on output rather than

                  insert blanks (at least after T is used to repeat cols.)


Unranked Status


        35        Allow additional format field descriptors, e.g., R And T.

        119       Specify NPEW.D output when N>D.


Not Recommended (Disapproved)


        17        Special format descriptor for vertical space control.

        47        Conditional compilation.

        53        Macro facility.

        58        Lower case alphabetics.

        65        Allow comments between continuation lines.

        71        Allow parentheses as field separators in formats.

        74        Blank input field detection.

        83        Complex format field descriptor.

        121       Pictorial format specifications,

        129       Allow alphanumeric stat. labels.




GROUP 2 - DATA TYPES, IDENTIFIERS



Approved


        2        Apostrophe as Hollerith delimiter.

        7        More than three dimensions.

        20       Generalization of subscript expressions.

        63       Delete second level definition.

        78       Allow subscript expressions less than one.

        124      Do not refer to a block name as a data name in std.

        132      Specify parts of complex no, in storage.

        165      Allow integer expressions in adj. dimensions.

        166      Allow vars. in common to be used for adj. dimensions.

        167      Allow Adj. Dimensions To Vary Within Subprogram

                 the adjustable array properties.


Must be in Standard


        6        Have Hollerith or character data type, allow Hollerith

                 constants in assignment and IF statements,


Should be in Standard


        45       Statement label data type.

        89       Binary or Boolean data type.

        128      Constant and input value of identical form must have

                                        same value.


Recommended


        154      Allow real constants to have more digits than a processor

                 can use.


Of Some Value


        34       Allow more precise precision requirements in

                 type-statements.

        39       Some standard way of packing and unpacking more than one

                 integer value in one storage unit.

        61       Structure data type.

        69       Allow two storage unit integers, esp. in common.


Of Trivial Value


        70        Specify some processor-defined constants, e.g. PI.

        81        Allow statement nos. as data, e.g. with trailing s.


Not Recommended (Disapproved)


        24        Names longer than six characters.

        52        Cluster data,

        153       Specify parts of double prec. no, in storage.



GROUP 3 - NONEXECUTABLE STATEMENTS, PROCEDURES AND SUBPROGRAMS,

                


Approved


        13        IMPLICIT statement.

        30        Expand basic external function list.

        164       Allow functions with no arguments.


Must be in Standard


        4         Multiple entry subprograms.

        77        Specify units of std. function args. and results.

        107       Allow naming of block data subprograms.

        108       Allow more than one block data subp. in one program.

        112       Allow equivalence to give two names to one variable.

        113       Allow subprograms to retain local vars. between refs.

        125       Combine basic ext. funs. and intrinsic funs.

        137       Allow common to be initialized in any program unit.


Should be in Standard


        8         Nonstandard return statement.

        11        Automatic typing of function names, i.e.,type of argument

                  determines which function is actually used.

        27        Implied DO loop in data statement.

        28        Array name without subscript in DATA statement.

        97        Allow function name in type-stat. within that fun. subp.

        126       Specify status of user subp. with name of basic ext. fun.

        131       Repeat more than one constant in DATA statement.


Recommended


        22        Internal subprograms.

        38        Parameter statement or equivalent.

        54        Storage classes.

        151       Make comma after slash in data stat. optional.

        152       Make comma between lists in EQUIVALENCE stat. optional.

        155       Define some intrinsic functions for some machine

                  dependent values.

        156       Define SIGNUM(X) function to produce +1.0 for X>0.,

                  0. for X=0., and -1.0 for X<0.

Of Some Value


        41        Standard overlay or segmentation scheme.

        67        Allow array name without subscript in equivalence.

        104       Allow no. of consts. to differ from no. of vars.

                  in DATA statement.

        130       Add more truncation and remaindering functions.

        134       Allow subp. refs. to use less than no. of dummy args.


Of Trivial Value


        12        Data in type-statements.

        66        Allow 'var=constant' in DATA stat.

        135       Allow namelist to be used for a list of actual arguments.


Unranked Status


        31        Add some basic subroutine names.

        162       Include inner product function.


Not Recommended (Disapproved)


        48        Debug facilities.

        76        Allow Initialization Of Blank Common.

        82        Dynamic storage allocation of arrays.

        105       Make mod functions always give non-negative results.

        111       Allow referencing beyond dummy argument array.

        114       Do not require labeled common blocks to be same length.

        160       Allow dummy arg. to be equivalenced to noncommon and

                  nondummy arg. array.


GROUP 4 - EXPRESSIONS


Approved


        29        Group successive exponentiations right to left.

        87        Allow array elements in statement function definition.

        148       Allow array arithmetic.


Must be in Standard


        1         Mixed real-integer arithmetic.

        90        Mixed real, integer, and double prec. arithmetic.


Should be in Standard


        123       Add < and > to character set and allow in rel. expr.

        149       Allow three new characters for .AND.,.OR., and .NOT.


Recommended


        49        Error conditions in computations.

        120       Allow control of rounding in fl. pt. operations.

        136       More permissive logical IF, e.g. IF(I.EQ.2.OR.3).

                  also more permissive logical expressions.


Unranked Status


        147       Allow subarray specification statement.

        159       Allow limited expressions in data statements.

        163       Specify sequence of evaluation of expressions.


Not Recommended (Disapproved)


        14        Implied multiplication after a right parenthesis.

        94        Allow matrix arithmetic.




GROUP 5 - EXECUTABLE STATEMENTS


Approved


        3         Multiple replacement assignment statement.

        32        Generalize DO statement, i.e., allow expressions with

                  zero and negative values.

        60        Not restrict strings in STOP and PAUSE to octal digits.

        80        Reduce restrictions on extended range.

        85        Make characters in STOP accessible.

        91        Allow all arith. types of assignment statements.

        92        Make comp. GOTO act like continue when var. out of range.

        102       Make comma and list optional in assigned goto.

        103       Allow integer expression in computed GOTO.

        106       Allow optional comma after stat. no. in DO stat.

        140       Allow branching into range of any unsatisfied DO.

        141       Allow real nos. as parameters in DO statement.

        

Must be in Standard


        138       Make end line imply a STOP in main program.

        139       Make end line imply a RETURN in subprogram.


Recommended


        15        Logical masking statements.


Unranked Status


        145       Define value of control var. after satisfying DO loop.


Not Recommended (Disapproved)


        46        Repeat statement.

        50        Error handling mechanisms.

        51        Multi-tasking facilities.

        62        Null statement labels. e.g. in IF statements.

        68        Allow assigned GOTO in diff. prog. unit than ASSIGN.

        79        Allow two stat. no. IF statement.

        88        Do not allow extended range.

        93        Execute DO loop once when m1>m2.

        110       Allow more than one stat. as part of logical IF.

        146       Provide for indicating abnormal termination of a program.

        157       One statement to interchange two values.



GROUP 6 - INPUT/OUTPUT EXCEPT FORMAT


Approved


        18        Error detection on reading.

        19        End file detection on reading.

        33        Allow variable format statement labels in i/o statements,

                  e.g., allow integer variable name that has been assigned

                  a statement label 'value' in an assign statement.

        56        Inclusion of READ, PRINT, and PUNCH statements.

        64        Allow expressions in output lists.

        73        Allow consecutive slashes to cause blank print lines.

        101       Rewind has no effect when unit is at initial position.

        142       Allow Hollerith constants in output lists with associated

                  format field descriptors.

        150       Allow more than one file on one i/o unit.


Must be in Standard


        40        Direct access i/o statements.


Should be in Standard


        9         REREAD or DECODE.

        10        ENCODE or equivalent.

        23        Free format input data, e.g., use a comma as field

                  delimiter.

        44        Open and close files.

        55        Block I/O transfer.

        99        Specify action for omitted format statement.


Recommended


        21        NAMELIST statement.

        36        Define std. input data sentinel and make available

                  number of numbers and/or number of lines read.

        72        Allow reading more chars. than written, get blanks.

        115       Define record length.


Of some value


        42        Ability to read tape records of unknown mode and length.

        143       Program interlocks.




Of Trivial Value


        122       Specify normal system i/o unit numbers.


Not Recommended (Disapproved)


        26        Allow Hollerith information in output lists without

                  associated format field descriptor.

        75        Negative i/o unit implies backward operation.

        116       Provide for multi-reel files.

        127       Get conversion when var. type and format disagree.

        158       Define initial position of i/o unit.




APPENDIX C              12, Pleasington Drive,

Bury, Lancs. BL8 2ET


Tel:  061 764 5168 (Home)

061 736 6541 X 213 (College)

Dear Mr. Gatehouse,


Thank you so much for your kind and prompt response to my request

for help.


Project/oral examiners familiar with FORTRAN and/or BASIC are

urgently needed by the City & Guilds of London Institute for the June

1972 examinations for the Basic Certificate in Computer Programming.

The need is for


2 examiner in Sutton Coldfield on 21 June

1 examiner in Brighton on 22 June

1 examiner in Warley on a date yet to be fixed between

21 June and 5 July

1 examiner in Wandsworth on a date yet to be fixed

between 21st June and 5th July.


Details on the PROJECT/ORAL EXAMINATION requirements of the

scheme are given on pp. 6 & 7 of the enclosed brochure.


Briefly a project/oral examiner is required to receive

and examine a written project, question the candidate on the project,

allocate an assessed mark to the project (a mark will have already

been given by the candidate's tutor), and assess the candidate's

ability to communicate on computer and programming topics.


On average a project/oral examiner is expected to deal

with no more than four projects. Examiners conduct the oral examinations

in pairs which are allocated in such a way that one of the two examiners

will have studied the candidate's project. Projects are sent to

examiners about two weeks in advance of the oral examination. It is

our usual practice to ensure that one member of each examining panel

(duo) will have taken part as an examiner in this scheme on at least one

previous occasion.


Anyone interested in this work and willing and able to take

part in the June 1972 examinations in one of the places listed above

is invited to contact Miss V. Hoban at City & Guilds of London Institute

76 Portland Place (just across the road from the meeting), London W1N 4AA,

Tel: 01 580 3050. He will be invited to complete a form for the

records of examiners and suitable arrangements for the examination will

be negotiated with him in good time before the due date. 


Once again. many many thanks for your help.


Yours sincerely,


Bernard D. Rowe.



B. D. Rowe (MIOM MBCS)

Chairman,

Moderating Committee for Examinations in Computer Subjects

City & Guilds of London Institute.