ISO/IEC JTC/SC22/WG5 N1860 Minutes of Meeting of ISO/IEC JTC1/SC22/WG5 Hosted by DIN, the German Member Body for JTC1/SC22 Leibniz Supercomputing Centre (LRZ), Garching/Munich, Germany June 27 - July 1, 2011 List of Participants: John Reid (JKR Associates, UK) convenor Dan Nagle (GMU, USA) PL22.3 chair Reinhold Bader (Leibniz Supercomputing Centre, Germany) Malcolm Cohen (NAG, UK) Rudolf Gabler (Ludwig Maximilian University, Germany) (Monday) Thomas Henlich (Siemens Energy, Germany / Gnu Fortran, Germany) Steve Lionel (Intel, USA) Bill Long (Cray, USA) Nick Maclaren (University of Cambridge, UK) Toon Moene (Gnu Fortran, Netherlands) David Muxworthy (BSI, UK) Rolf Rabenseifner (HLRS, Germany) (Tuesday) Craig Rasmussen (Los Alamos National Laboratory, USA) (Tuesday to Friday) Van Snyder (Caltech/JPL, USA) (Monday to Thursday) Masayuki Takata (Edogawa University, Japan) Wolfgang Walter (TU Dresden, Germany) (Wednesday and Thursday) Jim Xia (IBM, Canada) (Monday to Thursday) WG5 papers are referenced as Nnnnn. They are available from ftp://ftp.nag.co.uk/sc22wg5/ INCITS/PL22.3 is abbreviated throughout to PL22.3 and its papers are referenced as 11-nnn. They are available from http://www.j3-fortran.org/ 1. Opening of the Meeting The meeting opened at 09:00 on Monday 27 June 2011. 2. Opening business 2.1 Introductory remarks from the Convenor The convenor said that the main objectives of the meeting were to complete the technical report on further interoperability with C and to process interpretations for Fortran 2008. Developments on coarrays should have lower priority. 2.2 Welcome from the Host Reinhold Bader welcomed participants to LRZ. He was congratulated on the excellent facilities. 2.3 Local arrangements The host announced two additional events: a tour of the computer museum at the Military University on Monday evening and a tour of LRZ on Wednesday, which would be followed by a dinner for participants. 2.4 Appointments for this meeting The drafting committee would be Reinhold Bader, Toon Moene, David Muxworthy (chair), Van Snyder, Masayuki Takata and Jim Xia. David Muxworthy would act as secretary and John Reid as librarian. 2.5 Adoption of the Agenda (N1852) The preliminary agenda was adopted. 2.6 Approval of the Minutes of the Las Vegas 2010 Meeting (N1810) The minutes of the previous meeting were approved. 3. Matters arising from the minutes There were no items not otherwise on the agenda. 4. Status of Las Vegas 2010 Resolutions (N1811) With regard to resolution LV8 it had been proposed that the next meeting would be held in the last week of June 2012. The convenor would circulate the membership during the week to check that this date was generally acceptable. Otherwise there were no outstanding issues. 5. Reports 5.1 SC22 Matters (Convenor) The convenor drew attention to his report of the 2010 SC22 meeting (N1837). It was noted that part 3 of the Fortran standard (Conditional Compilation) had now been withdrawn. Part 2 (Varying Length Strings) was a potential candidate for deletion in the future. 5.2 National Activity Reports (Heads of Delegations) Canada: The SCC Languages Committee met every three months; IBM provided the Fortran representative. There was no Fortran committee as such. Germany: The DIN Fortran convenor would attend later in the week. There was a small number of members but they were quite active. Japan: Report is in N1855. Netherlands: There was no Fortran committee as such. Fortran standards matters were dealt with by the general programming languages committee. UK: Report is in N1857. US: All US business was transacted within PL22.3. The TAG dealt with administrative matters. 5.3 Report from Primary Development Body (INCITS/PL22.3) Resource had been concentrated on developing the TR on Interoperability. There had been very active email discussion. 5.4 Reports from other Development Bodies (Editors/Heads) Interoperability TR: Bill Long The latest draft of the TR (N1854) had been produced for this meeting. 5.5 Liaison Reports: INCITS/PL22.11 (C): Craig Rasmussen WG14 had welcomed an earlier copy of the Interoperability TR and were particularly interested in the examples. It was agreed that a copy of the latest draft should be sent to WG14 for comment. MPI: Craig Rasmussen No report, in view of presentation by Rolf Rabenseifner on Tuesday (see below). UPC: Dan Nagle No information was to hand. IFIP/WG2.5: Van Snyder There were to be meetings of WG2.5 in Boulder in August and in Spain in July 2012. WG2.5 influenced IEEE to convene a project to develop an interval arithmetic standard. OpenMp: Matthijs van Waveren No information was to hand. The convenor would contact Matthijs van Waveren to ask if he was willing to continue as liaison. Concern was expressed that OpenMp 3.1 showed lack of awareness of Fortran developments. WG23 (Vulnerabilities): Dan Nagle WG23 proposed to update their TR (TR 24772) in September and thereafter produce annual revisions. There would be language-specific annexes for Ada, C and SPARC; a potential draft for a Fortran annex had been prepared. This was followed by PL22.3 opening business. 6. Construct the PDTR draft TR on Further Interoperability with C There was a presentation by Bill Long on items on the draft TR (N1854) on which the PL22.3 subgroup wished guidance. The guidance was given in discussion without the need for votes. With regard to the comments on the TR which were noted in N1848, it was decided to delegate to the individuals concerned to pursue their views in subgroup during the meeting. [PL22.3 plenary and subgroup sessions occupied the rest of the day.] The meeting adjourned at 16:00 in order to visit the Computer Museum (2.3 above). ........................................................................... Tuesday Attention was drawn to the fact that withdrawal of part 2 of the Fortran standard (N1375) would remove five character functions (in subclause 3.7 of that standard) which were not provided in Fortran 2008. This was discussed later in the week. [It was item 11 on the agenda; thereafter item 11 matters were considered as appropriate throughout the meeting.] There was a discussion on whether WG5 should produce a standard on interfaces to certain mathematical functions, analogous to those for C (ISO/IEC 24747:2009) and C++ (ISO/IEC 29124:2010). The resulting straw vote was: 0 yes - 5 no - 8 undecided. This was item 10 on the agenda. There was a discussion on whether WG5 was in favour of there being a Fortran annex to the TR on vulnerabilities (TR 24772). WG23 would accept input on Fortran only if it had been approved by WG5. After consideration the straw vote was: 8 yes - 1 no - 4 undecided. It was decided to aim for incorporation of the annex in the September 2012 revision of the TR, which depended on WG23 adhering to their current schedule. Dan Nagle, John Reid, David Muxworthy and Makki Takata agreed to participate in the development. A mailing list would be set up. [This was item 9 on the agenda] Rolf Rabenseifner gave a presentation (N1862) on the MPI-3.0 Fortran interface, in particular detailing concerns about incompatibilities and optimizations. [The remainder of the day was occupied by PL22.3 plenary and subgroup sessions.] The meeting adjourned at 17:00. ........................................................................... Wednesday The proposals, in 11-193r1, to make changes to the coarray collective procedures in the draft coarray TS were discussed with a view to obtaining WG5's opinion. The debate was adjourned to Friday. The matter of the five character functions (above) was raised again and was referred to subgroup for a recommendation. The convenor drew attention to a draft of a new strategic plan with a view to consideration the following day. [PL22.3 plenary and subgroup sessions occupied the remainder of the day.] The meeting recessed at 17:15 for a tour of the LRZ computer facility. ........................................................................... Thursday The first draft of the new strategic plan for WG5, N1859, was discussed. The committee's main concerns were that this gave insufficient time for scrutiny of public comments on the interoperability TR and that the anticipated TS on further coarray features was proceeding without objectives or requirements being adequately specified. The plan was withdrawn for revision. [The remainder of the day was occupied by PL22.3 plenary and subgroup sessions.] A revision of the strategic plan, N1859, was discussed. This gave two alternatives for the schedule for the Interoperability TR according as whether an extension to the time limit could be obtained. Extensions were granted by SC22 in units of one year but it would be intended to use only four months. 7. Consider the technical content of the proposed TR on Further Coarray Features. There was a discussion on a draft requirements paper for the further coarray feature TS. A straw vote on whether the requirements should include parallel input/output went: 2 yes - 1 no - 12 undecided. The main discussion was on the need to specify objectives. The meeting recessed at 17:15. ........................................................................... Friday [PL22.3 plenary, subgroup and closing sessions occupied part of the day.] A draft of the resolutions (N1861) was discussed and minor changes made. A resolution on varying length strings was added: it was agreed that the convenor would write and make publicly accessible an updated version of the ISO_VARYING_STRING module. This would mean that if or when part 2 of the Fortran standard were withdrawn, those character functions which were in part 2 but not in part 1 would still be generally available. [During the discussion it was said that five character functions were affected. In fact there are six functions and three subroutines.] There was a discussion on the draft requirements document for the further coarray facility TS (N1863). This centred on the need for defined objectives for the project and the extent to which the content should be based on the current draft in N1858 which was in turn based on resolution LV5 from 2008. A straw vote on whether to restrict the content essentially to N1858 was: 0 yes - 7 no - 4 undecided. The approach outlined in N1863 was therefore abandoned in favour of a plan outlined in N1859 which allowed for a public trawl for requirements, followed by specification of aims and objectives and selection of features by PL22.3 and WG5. This, with time scales, is summarized in resolution G10 and section 3 of the final version of the new strategic plan (N1859). Changes were made to the draft resolutions to accommodate the above decision and further minor edits to the resolutions were suggested. Reinhold Bader gave a tutorial on coscalars, which was followed by discussion. The slides are in document N1864. 12. Closing Business 12.1 Future meetings The 2012 and 2013 meetings had been dealt with earlier in the meeting. 12.2 Any other business None was raised. 13. Adoption of Resolutions (N1861) Resolutions G1, G2, G3, G12 and G13 were approved by unanimous acclaim. Resolutions G4 to G11 were approved by unanimous consent. 14. Adjournment John Reid thanked the local hosts again for their excellent support during the week. The meeting closed at 16:05 on Friday, 1 July, 2011. ........................................................................... The following agenda items are recorded here for completeness. 8. Consider the Fortran defect reports (interpretations) in J3-006. No interpretations were ready for consideration by WG5 at this meeting. 9. Construct a draft Fortran annex for the TR on "Guidance to Avoiding Vulnerabilities in Programming Languages through Language Selection and Use". This item was considered on Tuesday (above). 10. Consider whether Fortran should have a math functions module, as an optional annex or an optional separate part. This item was considered on Tuesday (above). 11. WG5 Business and Strategic Plans 11.1 Goals for 2011-2014 These had been considered throughout the meeting, particularly with reference to the schedules in N1859.