From J.L.Schonfelder@liverpool.ac.uk Mon Mar 13 10:37:55 1995 Received: from mailhub.liverpool.ac.uk (mail.liv.ac.uk) by dkuug.dk with SMTP id AA15503 (5.65c8/IDA-1.4.4j for ); Mon, 13 Mar 1995 11:44:02 +0100 Received: from pop.liv.ac.uk by mail.liv.ac.uk with SMTP (PP); Mon, 13 Mar 1995 10:38:05 +0000 Date: Mon, 13 Mar 1995 10:37:55 GMT From: Lawrie Schonfelder Subject: strategic miscue To: SC22WG5@dkuug.dk Message-Id: Priority: Normal Mime-Version: 1.0 Content-Type: TEXT/PLAIN; CHARSET=US-ASCII X-Charset: ASCII X-Char-Esc: 29 Jerry has made a passionate plea for a cut down condition handling. While last year I would have been prepared to argue for a change in schedule, and did so, so as to allow some clearly important functionalities to be added to Fortran as soon as possible, I do not believe this is now a viable proposition. Last year and previously I believed that it was of critical importance for three functionalities to be added to Fortran 9x, viz., Condition handling, allocatable components, and parameterised derived types. I think it would have been reasonable up until last August to have voted that these functionalities be added, even if 9x were to become 97 rather than 95. However, having taken the decision that x=5 I see no way in which we can retain any credibility unless we actually achieve F95. As a result all of these functionalities become F2K candidates. Unless we have a votable document in Tokyo we cannot get F95. The only votable document we can have is the one we currently have. To try to add Jerry's proposal at this stage I believe will add at least 18months to the time to getting a standard, if not more. I think we should produce F95, as is, as quickly as possible. WE should also review our working mechanisms and procedures. It should be possible to get requirements set much more firmly and earlier than we have done this time. We should then manage the development activity so as to ensure we implement these requirements as early as we can and get the information about these into the public domain. We should reach 1999 with only editorial processes remaining to be completed. I dont think no conditions in F95 is quite as disasterous as Jerry, unless we are actually saying Fortran has lost the game as a general purpose scientific and engineering applications language. If we think it is already a niche market language confined to high performance numerical market then Jerry could be right. If we think it still has a general purpose role, which I think it does, then we have to look at a more rounded development. F2K in this case is not too late but it MUST BE 2000 for standard and implementation soon after not 200x with implementation 200x+4. -- 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