From J.L.Schonfelder@liverpool.ac.uk Wed Jun 22 03:15:58 1994 Received: from mailhub.liverpool.ac.uk (mail.liv.ac.uk) by dkuug.dk with SMTP id AA04187 (5.65c8/IDA-1.4.4j for ); Wed, 22 Jun 1994 11:16:11 +0200 Received: from liverpool.ac.uk by mailhub.liverpool.ac.uk with SMTP (PP) id <10538-0@mailhub.liverpool.ac.uk>; Wed, 22 Jun 1994 10:15:45 +0100 Date: Wed, 22 Jun 1994 10:15:58 PDT From: Lawrie Schonfelder Subject: Requirements for F96/F2001 To: SC22/WG5 members Message-Id: Priority: Normal Mime-Version: 1.0 Content-Type: TEXT/PLAIN; CHARSET=US-ASCII X-Charset: ASCII X-Char-Esc: 29 The next five messages are ascii versions of papers containing proposals that I believe implement important requirements for F96 or F2001. In general I strongly believe that all of these are things that need doing for for 2001 and there are some which I believe are crutial for F96 if Fortran is to remain a serious language into the next century. F90 was published in 1991 but production quality compilers are only now becoming sufficiently widespread for serious users to start contemplating using F90. This in spite of the first class demonstrator work done by Malcolm and the Nag compiler. We can therefore predict that F96 compilers will similarly not appear in significant numbers until, say, 1998/9; unless of course the extensions included are so minor as to be irrelevant to user facility improvement. The F2001 compilers can be expected then in 2004. With this timescale in mind it is my view as an end-user watcher (it is my job to be so as these are the people who pay my salary as a computing service director) that unless F96 goes a significant way to remove the more obvious holes left in F90 and the deficiencies which users are now hitting when they start using it, by the next millenium Fortran will have become an interesting historical study like latin or sandscrit. The language of widespread application use will be C++ or its derivatives. The proposals I am making in these following papers (printed versions have been sent to Jeanne and the two Davids for distribution) are all aimed at removing significant deficiencies from the current F90 language which make its use in practice cumbersome, confusing or in some cases impossible. The papers are proposals including what I believe are complete edits to implement the proposals. I also believe that all the proposals are mutually consistent. They could all be implemented together without causing conflicts. Nor do I believe that there are any conflicts with the proposals already made re object and type initialisation, or the continuing work on allocatable components. I therefore believe that the technical job of incorporating the edits to implement these in the revised document in not huge and should not cause any undue delay in the production of the F96 standard. I am therefore strongly recommending that most if not all of these extensions be included in F96. The extensions and their characterisation are: Parameterised Datatypes - cruitial for F96 Generic Dummy procedures - cruitial for F96 Constructor enhancement - cruitial for F96 Duplicate declaration of interfaces - desirable for F96, cruitial for F2001 Class of names clean-up - desirable for F96, cruitial for F2001 I also believe exception handling is cruitial for F96, but as I indicated in a previous mail I believe the proposal currently on the table is fatally flawed in a few key features. Along with Mike Delves I have been working on a proposal to modify the current proposal to solve these problems. I hope to get this complete and into the distribution before Edinburgh. Unfortunately I am unable to come to the meeting but I hope you will look seriously at these proposals. I do believe that Fortran's future as a serious scientific applications language is in the balance and these extensions all of which improve Fortran's regularity and add to its OOP capabilities are essential to tip the balance the rigth way. Best regards Lawrie -- Dr.J.L.Schonfelder Director, Computing Services Dept. The University of Liverpool, UK e-mail J.L.Schonfelder@liv.ac.uk phone: +44(51)794-3716 fax: +44(51)794-3759