Minutes of meeting held on 29 November 1982

Present:   C Burse            National Children's Bureau

J L Dyke           H R C

W W Flexner        Statistical Office of the United Nations (Retd)

B S Hersman        Abbey Consulting Associates

D J Holmes         Rolls Royce Ltd (Bristol)

C Lazou            ULCC

C K MacKinnon      UKAEA

B Meek             Queen Elizabeth College

J D Murchland

M Nunn             CCTA

R Palmer           LAMSAC

D M Vallance       Salford University (Secretary)

T L van Raalte     MOD

J D Wilson         Leicester University (Chairman)


Apologies were received from David Muxworthy (PLU, Edinburgh Univer-


2.      MINUTES OF PREVIOUS MEETING  [13 September 1982]

5(2)    'Haydn' should be 'Heyden'

5(3)    It should be noted that the '60%' is divided equally between

        the Specialist Group itself and the BCS Special Projects Fund

        which is [administered by] the Specialist Groups Board.


The Chairman has updated and enhanced the Fortran 77 booklist which

appears as Appendix A of these minutes. Members are asked to notify

any further additions or amendments to the Chairman.

The Secretary has not yet written to the ACM regarding the ForTech

newsletter. A number of ACM members have not yet received a personal

copy of the first issue of ForTech. A copy of the application form

for ForTech is included as Appendix B of these minutes. The Secretary

will enquire whether BCS FSG members are entitled to a subscription



The situation is unchanged from that reported in the previous minutes.

The Treasurer has asked the BCS to progress the matter further.

A request for funding for the Group for £350 has been submitted to the


Non-BCS FSG members are asked to remit a £2 fee for 1982-83 to the

Treasurer. A slip for this purpose is provided as the last page of

these minutes.


The Chairman has received a letter from the Editor of Datalink asking

for the co-operation of the Group in keeping Datalink informed of

events relating to Fortran.

The Specialist Groups' Board met on l7 September 1982.

(a)  There was concern expressed at the meeting at the high salary

advertised for the post of Specialist Groups' Board Assistant

Secretary for Groups and Branches. The advert asked for an

administrator who was not required to have specialist comput-

ing knowledge.

(b)  It was noted that there is a new edition of BCS Handbook No 7,

'Counsellors, Consultants, Expert Witnesses and Arbitrators'.

(c)  The Royal Charter was once again discussed. The document will

eventually appear in the BCS Journal.


Brian Meek reported on a proposal he had made to the BSI to enable a

British Standard for a programming language to be defined which was a

subset of any existing international language standard. Brian Meek

had studied BS0 and had noted that the conditions for the existence of

a British Standard, that refers to an existing international standard

all exist for the current ANSI Fortran 77 Standard. The issue at BSI

is the format of any possible British Standard. Such a Standard would

be an essential pre-requisite to persuading the DOI, CCTA, etc. to

adopting 'firming up' measures with the current Standard. The intend-

ing user of such a 'British Standard Conforming Compiler' would be

assured, for example, of a certain minimum set of exception handlers;

that a certain size of program would always be allowed; etc.

Members of the Group who wish to comment on this proposal are asked to

contact Brian Meek directly (Queen Elizabeth College).

No FSG member present at the meeting had yet received a copy of the

minutes of the November X3J3 meeting. The 'S6' document will no

longer be circulated outside X3J3, but it might possibly be merged

with the current Fortran 77 Standard to form a working document.


Two minor errors and corrections were noted as follows:

The Secretary will attempt to keep the compiler survey up to date.

Micro implementations of Fortran 77 will be included.

The possibility was discussed of preparing a more comprehensive book-

let using the results of benchmarks from a variety of sources.


The Chairman has available a Fortran 77 statement summary. It was

agreed that the Group would pay the X3J3 Service fee of $75 in order

that the Secretary could continue to receive X3J3 minutes.


The next meeting of the Group will take place at BCS Headquarters on

Monday 7 February 1983 at 10.45 a.m. The afternoon session (2.00

p.m.) will be devoted to a presentation by Ken Brodlie of Leicester

University on the implications for Fortran of the Graphical Kernel



Chris Lazou of ULCC gave a talk entitled "The CRAY-l as a Fortran Engine".

Notes on the talk appear as Appendix C of these minutes.


   Books on FORTRAN 77

All prices are approximate and refer to paperback versions unless stated



Language Textbooks

FORTRAN 77 - Harry Katzan Jr.

        Van Nostrand Reinhold (1978).

FORTRAN 77 Programming - W. S. Brainerd, C. H. Goldberg & J. L. Gross

        Harper & Row (1978).

Programming in Standard FORTRAN 77 - A. Balfour & D. Marwick

        Heinemann Educational Books (1979)        Price #5-50 (also in

        hardback) A comprehensive textbook aimed at both novice and

        experienced user: not aimed at any particular discipline. Contains the

        full language syntax in an appendix. The authors are in the Computer

        Science department of Heriot-Watt University, Edinburgh.

FORTRAN 77 featuring Structured Programming -        Loren P. Meissner & Elliott

        I. Organick - Addison-Wesley (1979)                Price #10.

        A comprehensive textbook aimed at both novice and experienced user. One

        of the authors (Meissner) is secretary of ANSI X3J3.

FORTRAN 77. Principles of Programming -              Jerrold L. Wagener

        John Wiley (1980)                                  Price #10-85

        The author is a member of ANSI X3J3.

Structured Fortran, a FORTRAN 77 Programming Course -     T. M. R. Ellis

University of Sheffield Computing Services (1980)         Price #2-60

Handbook designed to accompany a series of 10 video-tape lectures

produced at Sheffield University. The course is aimed at novice users

and covers most of the language; the rest being provided as an Appendix.

The full syntax diagrams are included as an Appendix. The book is a

transcript of the tapes plus examples, language variants etc.. A more

comprehensive textbook by the same author is available, see below.

A Structured Approach to FORTRAN 77 Programming -              T. M. R. Ellis

Addison-Wesley (1982)                                     Price #8-95

Based on the earlier video-tape course this book is, however, a much

extended and revised work. The book provides a complete introduction to

programming in the Fortran language. Each Chapter concludes with a

summary of the main points and new language features introduced,

together with many examples. Technicalities, both hardware and software,

are avoided in the early Chapters. Emphasis is placed on good style and

systematic testing. The full language syntax is given as an Appendix.

FORTRAN 77 for Humans - R. Page & R. Didday

        West Publishing Co. (1980)

Programming with FORTRAN 77 -        J. Ashcroft, R. H. Eldridge,

R. W. Paulson & C. A. Wilson -        Granada Publishing (1981)

Price #5-95

A textbook for Mathematicians and Scientists. The authors are in the

Mathematics department of Salford University and the book is written in

the style of a mathematics textbook.

Programming FORTRAN 77 -        J. M. P. Hume & Holt

        Reston                                                  Price #8-40

FORTRAN 77 - D. M. Monro

        Arnold                                                  Price #8-50

Application_or Method Oriented Books

Fortran with Style, Programming Proverbs -                Henry F. Ledgard &

        Louis J. Chmura Jr. -                Hayden (1978)      Price #6.

Programming in Fortran, Structured Programming with FORTRAN IV & FORTRAN 77

        Vladimir Zwass - Barnes and Noble Books (1981)         Price #3-95

Business Programming in FORTRAN IV & ANSI FORTRAN 77, a Structured Approach

Asad Khailany -                 Prentice Hall (1981)      Price #9-75

Aimed at undergraduate business studies students, it provides basic data

processing and computer organisation material plus considerable detail

about the Fortran language. Contains useful chapter on program and

system design and documentation. Simple example programs illustrate

language features and form useful models for the student to taylor to

his own needs. Fortran is not justified over Cobol. Numerical business

applications (financial modelling, critical path analysis, linear

programming etc.) not covered. The book is up to date in detail and

approach. (A. Clarke Sept 1982)

Problem Solving and Structured Programming in Fortran (2nd. Edition) -

     F. Friedman & E. Koffmsn

     Addison Wesley (1981)

     New edition compatible with FORTRAN 77. Intended to provide more than

     adequate coverage for an introductory course.

[The somewhat random layout of this appendix reflects that of the original





11 WEST 42nd STREET, NEW YORK, NY 10036

    Please enroll me as a member of


the SIGPLAN Technical Committee on Fortran

Membership includes Newsletter subscription. Please make checks payable to

ACM, Inc., P.O. Box 12115, Church Street Station, New York, N. Y. 10249

-----------------------------------------------------------------------------------------  Annual Membership Dues are

Name  Please print or type                                                    $ 6  for ACM Members

                                                                           $ 3  for ACM Student Members

                                                                           $16 for non-ACM Members


Mailing Address                                                           ___ ACM MEMBER

ACM Member No. ---------

(see note regarding dues payment)


                                                                                   ___ ACM STUDENT MEMBER

ACM Student Member No. ---------

(see note regarding dues payment)


City                                State                                 Zip       


Enclosed annual dues $16





----- New address.  Please change my ACM record                          Enclosed is payment for annual sub-

scription of $16

Note Subscription is included with

membership dues

Note:       ACM members who expect to renew their membership within the next

six months should send no FORTEC dues now as they will be billed on

their renewal notice.  Those expecting to renew in seven or more

months should send one half the annual FORTEC dues now.

---- Please send information on ACM




    Chris Lazou, ULCC

The ULCC Cray-1 is configured as follows:


1 Mword (64-bit words; 8 x 8-bit characters) memory

MCU consisting of

-        Data General Eclipse

-        80 MByte Removal disc

-        Magnetic tape

-        Printer

-        2 VDUs

3 disc controllers

9 x 600M Byte disc drive

The Cray-1 is connected to a 16M Byte Arndahl [sic] V8

The Cray-1 is seen by users as a 'Fortran Engine', although a PASCAL

compiler will soon be available.

Parallelism is achieved by:

-  using independent functional units

-  using results from one Vector register as an operand to

   another function during the same clock pulse

Although the Cray clock period is 2.2 times faster than a 7600, the

average instruction time is only 1.5 times faster.

The Cray Fortran compiler is referred to as CFT.

Cray state that CFT conforms to the ANSI Fortran 77 Standard. The

compiler has passed the FCVS validation tests. There are, however, a

number of problems with the latest release.

Not all the Salford Fortran 77 compiler test programs work correctly.

Length of a CHARACTER variable or constant is limited to 255 charac-

ters. The construct

        CHARACTER X*255

        X(32:33) = 'AB'

is not always recognised.

There are problems with CHARACTER statement Functions.

Output using 'TR500, TL499', which should be equivalent to 'TR1', does

not work.

CFT is a 2-pass compiler. The first pass builds intermediate forms on

a file. The second assigns registers and generates code. The com-

piler is written in assembler.

The compiler control directives provide options for source file input,

listing output, binary file output and for compiler options. There is

a specific option to control the number of 'vectorisations inhibited'

messages that are output. it is possible to obtain the assembly

language (CAL) listing of a program. A 'double-real' option is

provided to allow single precision programs to be compiled

automatically in double precision. There is a facility for 'flow-

trace' at run time to help with hand optimisation.

There is no ANSI compile-time option at present. Array bound checking

is available, as is Postmortem Dump.

Language extensions are allowed:

- Names can consist of up to 8 characters

- Lower case and upper case letters can be used interchangeably

- CFT allows the " character in addition to the Standard character set

- Comments are allowed of the form:

                A = B+C ! This is a comment

The compilation system provides a scientific library which includes

LINPAC and EISPAC in addition to the normal intrinsic routines.

Dates, times etc are available from a systems library.

The following data types are provided:

24- and 64-bit integers

        64-bit real

128-bit (software) double precision

2 x 64-bit complex

The 24-bit integers are obtained by means of a compiler directive.

Boolean constants (bit constants) are available.

Logical .TRUE. is represented by a positive value and .FALSE. by a

negative value.

Apostrophes (') or quotation marks (") can be used for character


.NOT. , .AND. and .OR. may be abbreviated to, respectively, .N.

.A. and .O. .XOR. (.X.) is available as a language extension.

NAMELIST and POINTER are available as extensions. NAMELIST is

implemented as in CDC and IBM Fortran. The syntax of the POINTER

statement is

                POINTER (P, A)

where P is a scalar and A is either a scalar or an array.

There is :a limit of 26 unnamed BLOCK DATA subprograms. These are

given different local names by CFT only if they appear in the same


The benefits of vectorisation are not significant until arrays become

greater than 4 or 5 elements. The compiler will only vectorise every

program; indeed, some programs to not lend themselves to vec-


The manual is perhaps too brief and lacks examples. 'Boundary con-

ditions' are not always stated explicitly.


I am a non-BCS member and wish to remain a member of the BCS Fortran

Specialist Group for the period April 1982 to March 1983. I enclose a

cheque value £2, payable to "BCS Fortran Specialist Group".


Organisation (if applicable):


Please send this form, together with your remittance, to

Mr T L Van Raalte


AWRE Aldermaston


Berks RG7 4PR