From jwagener@amoco.com Tue Apr 5 08:36:29 1994 Received: from interlock.amoco.com by dkuug.dk with SMTP id AA03231 (5.65c8/IDA-1.4.4j for ); Tue, 5 Apr 1994 20:27:37 +0200 Received: by interlock.amoco.com id AA12246 (InterLock SMTP Gateway 1.1 for sc22wg5@dkuug.dk); Tue, 5 Apr 1994 13:27:25 -0500 Received: by interlock.amoco.com (Internal Mail Agent-3); Tue, 5 Apr 1994 13:27:25 -0500 Received: by interlock.amoco.com (Internal Mail Agent-2); Tue, 5 Apr 1994 13:27:25 -0500 Received: by interlock.amoco.com (Internal Mail Agent-1); Tue, 5 Apr 1994 13:27:25 -0500 From: jwagener@amoco.com X-Openmail-Hops: 1 Date: Tue, 5 Apr 94 13:36:29 -0500 Message-Id: Subject: International Standards, Global Trade, and F77 To: sc22wg5@dkuug.dk X-Charset: ASCII X-Char-Esc: 29 Item Subject: Message text You will be interested in two meetings that I had in Washington DC last week. The first one was on March 30 at the National Academy of Sciences and the other on March 31 with the X3 secretariat. I'll briefly summarized the essence of each meeting and follow that with a document, that I whipped together following those two meetings, for possible action at Tahoe or at least no later than the Edinburgh X3J3 meeting. First the meeting on March 30. I was asked to represent Amoco's chief scientist at a conference sponsored by the National Research Council's project on "International Standards, Conformity Assessment, and US Trade Policy". (They wanted someone with standards experience, so I was "it".) The project is related to improving US competitiveness in global trade, not an area that I claim to know much about. From the distributed material I couldn't tell exactly what the conference was to be all about. But was it ever interesting and eye opening; and it turned out that I was the right guy for Amoco to send. You will no doubt note a touch of excitement in my tone, and a resulting strongly-worded (possibly ill-advised) document. The conference turned out to be a love-in for international standards; the villians were national (or regional or other forms of provincial) standards. It would appear that the current conventional wisdom, at US policy levels, is that private, voluntary international standards signifcantly help companies, and consequently nations, to compete effectively globally and that national (and especially government) standards significantly hinder such competitiveness. 'Course I knew it all along (:-). Actually, despite my long experience in and commitment to international standards, until this conference I had not truly assimilated the understanding of how non-international standards can seriously impact the ability of companies like Amoco to compete in international markets, and hence how provincial standards can adversely affect the nation's economy. Well, now that light burns brightly for me, and it feels good. About 40 of the roughly 150 particpants in the conference were from the US government, and it was amusing to see some of them wasting our time with their mumbo-jumbo bureaucratic positioning. (At one point the person sitting next to me and I turned to each other and whispered in unison "bureaucrat talking to bureaucrat" - we hardly understood a word of it.) But one of the highlights of the meeting (for me, at least) was a crisp, articulate description by an under-secretary of the DOD of, and I quote, "... [the DOD's] total reengineering of the acquisition process ... [and to] encourage contractors to find better alternatives to mil-specs and mil-standards". I still can hardly believe it - the US economy must be in worse shape than I thought. It remains to be seen, of course, if this will all come about, but I certainly take it as a harbinger of spring time for international standards and autumn for provincial ones. Finally, the emphasis was clearly on private, voluntary, bottom-up standards, such as the one we are working on in WG5 and X3J3. But then we all knew we were on the right track all along, right? So much for day one. On March 31 I visited the X3 secretariat, and first of all briefed them on the preceding day (none of them had been invited - only the higher-up CBEMA management had been invited). The other two things we discussed were the X3J3 annual report (which I'm currently working on and which I have to present to the OMC on May 5 - I'll have to play hooky from the X3J3 that morning), and the recent letter ballot we had on the Fortran 77 issue. It turns out, contrary to my earlier impression (groan), that our work is not quite done on the latter - we have to (groan again) respond to the negatives and submit those responses with the votes to the OMC. I want the committee to spend a minimum amount of time on this (ha, I know, wishful thinking) so I have drafted potential responses to all the points raised in the ballot comments. That's the following document. The vote of the membership was to recommend withdrawal of F77 and, spurred on by the euphoria of the March 30 meeting (and not taking time to count to ten), I have worded the responses so that they are not wishy-washy and that there is no doubt what the X3J3 recommendation is. Of course, even though you haven't seen the ballot comments yet, you need to keep me honest and see that I've touched all the necessary bases and possibly help me rein in the rhetoric a bit. Then it'll be up to the Tahoe meeting attendees whether we should vote on these responses or wait until the Edinburgh meeting to vote. (Note however that this note makes the two-week rule in case we want to vote at Tahoe.) In any event, the process seems now to be such that there is no way the X3-and-above-level action on this issue will be completed before mid-1995. So those wanting to delay the withdrawal until 1995 will get their wish. And in the meantime I still hope to minimize the time that the committee need spend on this issue. With all that, here are the responses I've drafted. Please send me any suggestions for changes, additions, etc., and I'll bring a revised version, with a document number, to the Tahoe meeting next month. Jerry ---------------------------------------- The complexity issue -------------------- The argument that Fortran 90 is too complex and will kill Fortran is an old one. X3J3 emphatically believes that this view is incorrect, but in the unlikely event it turns out to be correct, and since Fortran 90 is the (only) international Fortran standard, the proper remedy is to fix the international standard, not prolong an out-dated national standard. The timing issue ---------------- X3J3 concurs with the suggestion that the Fortran 77 standard be withdrawn in 1995, not in 1994. The FIPS issue -------------- X3J3 understands and respects a vote to not withdraw the Fortran 77 standard based on the FIPS situation. (Fortran 77 is a FIPS standard whereas Fortran 90 is not.) X3J3 understands that some US government procurement policies require FIPS conformance. However, X3J3 believes that itOs the FIPS that is broken in this case and that the Fortran 77 withdrawal 2-by-4 might be whatOs needed to initiate repairs. There are two reasons why this position might be less outrageous than it might at first seem. First, X3J3 notes that the DOD has publicly announced that it is drastically reengineering its procurement procedures and, to the extent allowed by law, encourage vendors to use alternatives to government standards where that makes technical and economic sense. In these challenging economic times the deficiencies in government procurement policies are all too clear, and in this light it does not seem likely to X3J3 that in practice an out-dated FIPS will be a serious stumbling block to reasonable and advantageous use of technology alternatives in government procurements. Second, Fortran 77 is a subset of Fortran 90, and any technical reason for the Fortran 77 FIPS can be satisfied by corresponding specifications involving Fortran 90. It would appear to X3J3 that this would formally allow continued use of existing Fortran 77 products for procurements that need only this subset of Fortran 90, though as noted above X3J3 believes that under current conditions such formality is moot. The validation issue -------------------- While there is no official validation suite for Fortran 90, X3J3 notes that there are at least two commercially available Fortran 90 validation suites. Though neither of these is OofficialO, there has been considerable effort to make them robust and comprehensive. X3J3 believes that in view of the existence of these validation suites, neglect by the responsible bodies to establish an official validation suite for the international Fortran standard should not be allowed to prevent withdrawal of an out-dated and redundant national standard. The stability issue ------------------- There are still unresolved interpretation requests for Fortran 90, as there were for Fortran 77 well past the mid-1980s. X3J3 does not believe that any of these requests are sufficient Oshow stoppersO to justify delaying withdrawal of Fortran 77 past 1995. The maturity issue ------------------ With respect to the argument that Fortran 90 compilers are not yet adequately available, X3J3 notes that Fortran 90 compilers are commercially available for computing environments from PCs to unix workstations to high performance architectures, and that there are examples of large organizations, including multi-national companies, using Fortran 90 for production applications. As the saying goes in fast-moving technical arenas, Oif you wait for things to settle down youOll never do anythingO. X3J3 believes it is time to complete the replacement of Fortran 77 with Fortran 90. ---------------------------------------------