From LJM@SLACVM.BITNET Wed Jun 30 03:43:00 1993 Received: from vm.uni-c.dk by dkuug.dk with SMTP id AA26117 (5.65c8/IDA-1.4.4j for ); Wed, 30 Jun 1993 21:14:40 +0200 Message-Id: <199306301914.AA26117@dkuug.dk> Received: from vm.uni-c.dk by vm.uni-c.dk (IBM VM SMTP V2R2) with BSMTP id 4740; Wed, 30 Jun 93 21:15:18 DNT Received: from SLACVM.SLAC.STANFORD.EDU by vm.uni-c.dk (Mailer R2.07) with BSMTP id 9897; Wed, 30 Jun 93 21:15:17 DNT Received: by SLACVM (Mailer R2.08 R208004) id 5567; Wed, 30 Jun 93 12:05:24 PST Date: Wed, 30 Jun 1993 11:43 -0800 (PST) From: "Len Moss" To: "SC22/WG5 Mailing List" Subject: SLAC's comments on Draft Final Report of ANS/IEEE 1003.9 WG X-Charset: ASCII X-Char-Esc: 29 For your information, here is a letter which I just posted on behalf of SLAC to Michael Hannah and the POSIX.9 WG mailing list. In case you haven't seen the report referenced in the subject line, it proposes to abandon the current project to produce a Fortran 90 binding to POSIX (I believe this report has been distributed fairly widely, but if you'd like a copy let me know). The POSIX.9 working group is meeting on 12 July 1993 and is expected to complete work on this report and submit it to the POSIX executive committee. Although this is probably too soon for WG5 or any of the national Fortran committees to prepare a response, I'd like to encourage members of these groups to send their own comments to Mr. Hannah (mjhanna@sandia.gov) and the POSIX.9 mailing list (posix-fortran@sandia.gov). - - - - - - - - - - - S L A C L e t t e r - - - - - - - - - - - - Michael J. Hannah Chair ANS/IEEE 1003.9 (Fortran Bindings) mjhanna@sandia.gov Dear Mr. Hannah: We at the Stanford Linear Accelerator Center (SLAC) are disappointed at the lack of interest that has been shown so far in POSIX.19, the project to produce a POSIX/Fortran 90 binding, and we are concerned to learn that the POSIX.9 working group is planning to abandon this effort. We urge the working group to reconsider. SLAC is a national laboratory whose mission includes experimental research in several areas of physics, from elementary particles to synchrotron radiation experiments. Historically these fields have relied heavily on large Fortran codes running on mainframe and mini computers under operating systems other than UNIX. However the changing economics of computing are rapidly driving the lab to embrace open systems. Major new projects within the lab, such as the proposed B Factory, are critically dependent on the success of this migration. With our many years of Fortran experience and our large inventory of Fortran software, a Fortran binding to POSIX is essential to us. The recently approved Fortran 77 binding is awkward and unnatural to use due to the limitations of Fortran 77, and we believe it will be inadequate to our needs. Until now, SLAC has not stepped forward to join the POSIX.9 working group due to our lack of UNIX experience. We are, however, represented on X3J3 and expected to make a contribution through its liaison activities with POSIX.9 (the Fortran side is, after all, where we have more expertise). We are now reevaluating our commitments to standards work and will consider becoming directly involved in the POSIX.19 project. Perhaps the proposal to terminate this project will convince a few other organizations facing a migration to workstations to do the same. We suspect, in fact, that the groups most in need of this binding tend to be new to the UNIX world and thus most hesitant to get involved. The draft final report of the working group mentions "an insufficient body of knowledge, practice, or users of these new features of Fortran 90" to resolve a number of open technical issues. While it is true that native Fortran compilers are not yet available we believe they will arrive soon. High quality translators have been available for some time, and these should be adequate to resolve some of the technical issues raised in the draft report (in particular, issues 5 and 6, regarding null-terminated strings and structure handles). Many of the other issues could best be resolved through liaison activities with X3J3. The draft report also suggests that "it is more appropriate for language extensions to be sponsored by the language committee" and quotes an ISO technical report, ISO/IEC TR 10182 to support this view. However, the international Fortran committee, JTC1/SC22/WG5, appears to disagree. WG5 is in the process of drafting a "Guideline For Binding To Fortran 90" as its response to TR 10182. The current draft of this document (ISO/IEC JTC1/SC22 WG5-N889, 19 May 93) contains the following recommendation: WG5 recommends...that in general the committee responsible for the functional specification is better placed to define a functional binding to Fortran 90 than is WG5...However WG5, in accordance with [TR 10182] guideline 3 [dealing with language extensions], should be involved with consultation on Fortran 90 bindings as early in the design process as possible... In addition, the Fortran committees are currently committed to an aggressive timetable for producing two new revisions of the Fortran language by 2000. In any case, as presently constituted they lack the necessary UNIX expertise to draft this binding themselves and would find it no easier than the POSIX community to recruit the necessary subgroup of experts. Instead of abandoning the project, we suggest that the POSIX.9 working group seek closer ties with X3J3 in order to call on their Fortran expertise. For example, why not submit the list of technical issues in the draft final report to X3J3 and request guidance? A busy technical committee is much more likely to respond to specific questions than to take on a whole new standard-writing project. Moreover, this is an excellent way to recruit more members for the POSIX.9 working group. In our opinion, abandoning the POSIX.19 project at this time would do a disservice to both the Fortran and the POSIX communities. Sincerely, Leonard J. Moss SLAC Representative on X3J3 ljm@slac.stanford.edu Charles R. Dickens Director, SLAC Computing Services dickens@slac.stanford.edu Adam M. Boyarski B Factory Computing amb@slac.stanford.edu -- Leonard J. Moss | My views don't necessarily Stanford Linear Accelerator Center, MS 97 | reflect those of SLAC, P.O. Box 4349; Stanford, CA 94309 | Stanford or the DOE