Title: Letters to the Editor

I am writing to you as Chairman of the British Computer Society Fortran Specialist Group as well as a member of the ACCU of 10 years standing.

I would like to expand on your mention of restrictions on coding style and formatting imposed by Fortran 77 in Pete Goodliffe's article in C Vu Volume16 No 4 and to let the readers of C Vu know that Fortran is alive and well in the 21st century.

I spent 15 years writing and maintaining Fortran 66 and 77 code. While the fixed format source form put some restrictions on the layout of program statements, they had to be within columns 7 to 72 of each line, there was no restriction on using indentation to show code structure and you could continue long statements over several lines. If I remember correctly the standard allowed for 9 continuation lines but many compilers allowed up to 99.

The next revision of the ISO standard, Fortran 90, introduced the free format source form, where there was no restriction on the positioning of statements on a line other than a maximum line length of 132 characters, which some compilers increased. It introduced more modern features to complement Fortran's well known strengths in numerical computation. Fortran 90 and 95 introduced operations which could be carried out on whole arrays or sections of arrays, rather than just on individual elements. Also dynamic memory allocation and abstract data types were introduced.

The most recent version of the language, Fortran 2003, is due to be published in the autumn of 2004 and contains features to enable object orientated programming to be carried out in Fortran. For more information on the development of Fortran standards since Fortran 95 please see the Standardisation page on the Group's web site at http://www.fortran.bcs.org/standards/stanhome.htm.

If anyone is interested in exploring the modern features of Fortran a version known as F has been developed. F is a subset of Fortran 90/95 that enforces correct coding practices by removing antiquated and dangerous features in F90/95. There are new versions for Linux, Solaris, and Windows available for free download from http://www.fortran.com.

In response to some of the questions posed in Pete's article I can say that I tried to code in a consistent manner when writing new code, using 2 spaces for each level of indentation in both Fortran and C, and to "improve" the layout and structure of existing Fortran code when I had to modify it and had the time for cosmetic changes!

From my own experience I agree with Pete that tabs should not be used for indenting. We were programming across several platforms, each of which had its own editor, which handled tabs differently so that tab indented code could look OK in one editor but be almost unreadable in another. I aimed to globally replace all tabs with 6 or 8 spaces whenever I came to work on a tab-indented file. This was possible because we were only a small team, three to five developers, and we each tended to work on a particular area of the code so formatting changes did not often get changed back by someone else!

I should like to take this opportunity to say how much I have enjoyed Pete's articles on Professionalism in Programming over the last four years. I have found relevant and informative points in every one.

While writing about professionalism I would like to remind members of the ACCU that the British Computer Society undertook a major relaunch earlier this year using the slogan "Making IT the profession for the 21st century" and aimed at making individual membership more relevant to professionals in IT. See http://www.bcs.org for more information.

Peter Crouch <pccrouch@bcs.org.uk>