From jwagener@amoco.com Wed May 10 08:35:20 1995 Received: from interlock.amoco.com by dkuug.dk with SMTP id AA18671 (5.65c8/IDA-1.4.4j for ); Wed, 10 May 1995 20:36:20 +0200 Received: by interlock.amoco.com id AA02970 (InterLock SMTP Gateway 3.0 for sc22wg5@dkuug.dk); Wed, 10 May 1995 13:35:59 -0500 Message-Id: <199505101835.AA02970@interlock.amoco.com> Received: by interlock.amoco.com (Protected-side Proxy Mail Agent-3); Wed, 10 May 1995 13:35:59 -0500 Received: by interlock.amoco.com (Protected-side Proxy Mail Agent-2); Wed, 10 May 1995 13:35:59 -0500 Received: by interlock.amoco.com (Protected-side Proxy Mail Agent-1); Wed, 10 May 1995 13:35:59 -0500 From: jwagener@amoco.com X-Openmail-Hops: 1 Date: Wed, 10 May 95 13:35:20 -0500 Subject: informal JLW reports on the April'95 WG5 and X3J3 meetings To: sc22wg5@dkuug.dk X-Charset: ASCII X-Char-Esc: 29 Item Subject: Message text WG5 meeting, 1995 Apr 17-21, Tokyo ---------------------------------- In addition to the Japanese host delegation (6 members) and the convenor (UK), there were three other delegations: Germany (1 member), the UK (4 members), and the US (2 members). The objectives of the meeting were to make final plans for the Fortran 95 CD ballot and to make initial plans for Fortran 2000 requirements. Germany came to the meeting prepared to delay the CD for up to a year to allow floating point exception handling and some interoperability with C to be accommodated in this revision of the Fortran standard. The other delegations all wanted to push forward with the current draft, X3J3/95-007 (though individual members of some of these delegations may have agreed with the German position). In the end the decision was to submit a revision of X3J3/95-007 for CD ballot next month, the revisions specified in document N1112 generated at the meeting (N1112 contains no Fortran 95 requirements for exception handling and interoperability). However, WG5 noted that several items that didn't make it into Fortran 95 are too critical to delay until Fortran 2000; these include the two Fortran 95 requirements that X3J3 could not complete in the established timeframe: exception handling and allocatable components. Interoperability with C seems to have widely become considered to be of increasing criticality and was added, without dissent, to the list of critical items. In addition, parameterized derived types where strongly argued to be critical and were consequently also added to the list. To address these critical items, WG5 at this meeting devised a "technical report strategy" to effectively standardize, and hence encourage implementation of, these critical items well before Fortran 2000, but in such a way that they could be incorporated into Fortran 2000 without change. The mechanism is the ISO type 2 technical report, and document N1111 describes how this mechanism will be used to address these critical items with the desired outcome. Consequently, three technical reports were defined and initiated: one for floating point exception handling (the most critical subset of exception handling - the technical reports are to be used for relatively small, self-contained items) - Wolfgang Walter interim project editor, one for interoperability with C - Jamie Shiers interim project editor, and one dealing with data type enhancements that encompasses both allocatable components and parameterized derived types - Lawrie Schonfelder interim project editor. Other than the topics for the technical reports, no further work was done at this meeting relative to Fortran 2000 requirements. Jerry Wagener ====================================================================== X3J3 meeting, 1995 Apr 24-28, Maui ---------------------------------- 19 of the 22 current X3J3 members were present or represented at this meeting. The primary objective of the meeting was to revise 95-007 in accordance with N1112, to prepare 95-007r1 (the CD ballot document) by the end of May 1995. Secondary objectives were to continue work on the unresolved interpretations and to respond to the WG5 resolutions of the previous week. Most of the committee effort during the week was toward the primary objective. By the end of the week all items related to revision of 95-007 were successfully resolved; thus 95-007r1 will contain no deficiencies in terms of the requirements of N1112 (an important part of which was X3J3 document 95-103), and 95-007r1 will be available on schedule for the CD ballot. At the beginning of the meeting there were 27 interpretation items needing X3J3 attention; 9 of these were resolved during the week and will be letter balloted by the committee. One subgroup (/jor) was charged during the week to respond to the WG5 resolutions and in particular to draft the X3J3 strategy relative to the technical report process established the preceding week. Quite a bit of full committee time was devoted to discussing the WG5 action on technical reports and on discussing and voting on the /jor recommendations. Document 95-005r1 (Responses to WG5 Resolutions) was approved by unanimous consent and document 95-128r1 (detailed response to resolution T7, which deals with the technical report process) was approved by a vote of 14-1. Document 95-128r1 details the problems that X3J3 sees with the technical report process, but pledges X3J3 support in developing the technical reports and describes how X3J3 proposes, as the primary development body for Fortran 2000, to interact with the technical report development bodies. X3J3 concurs with the list of critical items that WG5 has identified for technical reports, with the exception of parameterized derived types. As stated in the X3J3 response to resolution T11 (data type enhancements), X3J3 believes that parameterized derived types does not meet the criteria for technical report items as described in N1111 and therefore requests WG5 to remove parameterized derived types from this technical report. Finally, the committee approved a letter (95-145r2) to be used in announcing the US public review associated with the upcoming CD ballot. Jerry Wagener ======================================================================