Minutes of Atlantic UBL TC call 20 April 2005

[jon.bosak@sun.com]

MINUTES OF ATLANTIC UBL TC MEETING
15H00 - 17H00 UTC WEDNESDAY 20 APRIL 2005

ATTENDANCE

   Jon Bosak (chair)
   Tony Coates
   Mavis Cournane
   Mark Crawford (vice chair)
   Stephen Green
   Mike Grimley
   Anne Hendry
   David Kruppke
   Paul Thorpe

STANDING ITEMS

   Additions to the calendar (http://ibiblio.org/bosak/ubl/calendar.htm)

      None.

   Liaison reports

      None.

   Subcommittee report: SSC

      Has not been meeting while discussing [ver] and will not
      meet tomorrow.

   Subcommittee report: SBSC

      StephenG: Have just done another pass at index.htm and sent
      out to the SC, but this is not finished yet.  It includes an
      example of BPSS referencing the SBS.

      JonB: Let me know when you want me to make another edit.

   Team report: ver

      StephenG: The team came up with two alternatives.  (1) The
      default: redeclare everything in a new set of schemas and
      remove NDRs VER8 and VER9; or (2) move towards polymorphic
      processing support for both major and minor versions, which
      would require making everything global (thus allowing
      customizers to use the polymorphic approach even if we
      don't) and allowing a limited use of substitution groups in
      the construction of minor versions.  The recommendation is
      to make the next version a major version with every element
      declared global; this wouldn't commit us to one method or
      the other.

      MarkC: See my mail to the list a couple of hours ago.  The
      easiest thing is to make the next version 2.0, but this
      seems to abandon our goal of allowing pieces to be updated
      without requiring monolithic releases.  Changing everything
      to global is not a problem.

      JonB: We could make it clear that only this release is
      monolithic, and that we're doing this to give us room to put
      in place a workable minor versioning strategy that we can
      live with going forward.

      MarkC: It would be good to align with ATG on the CCTS
      modules.

      MikeG: Agree with MarkC: don't want to lose the original
      intent of ModNamVer, but could do for this time if we make
      it clear that we do not intend to make monolithic releases
      the norm.  It would allow us to align with the ATG schemas.

      AnneH: It's probably best to call this 2.0 and try to get in
      as many known backwards-incompatible changes as possible.

      DavidK: Think it's OK to go to 2.0 because there are some
      errors in 1.0 that would require such changes anyway.

      MavisC: OK with the second option.

      PaulT: No opinion.

      JonB: Should we say while we're working on it that the
      limited substitution-group approach is the one we're
      pursuing for minor revisions?

      MarkC: Still not convinced that there is no alternative to
      substitution groups, and we still need to deal with
      namespace versioning.  Don't think we need to get rid of
      VER8 and VER9.  In addition to global/local we would still
      need to discuss token [namespace prefix] versioning: NMS 8,
      10, 12, 14, 16.  Are we going to adopt the ATG schemas as
      they are, or are we going to publish them with
      modifications?

      AGREED:

       - We will call the next version UBL 2.0

       - We will redeclare the whole schema set

       - We will explain that we are not adopting this monolithic
         approach as our strategy for versioning and that we
         still intend to produce a methodology for minor versions

       - We will make everything global (i.e., change ID and Code
         to global)

       - We will align with the ATG NDRs as far as possible, in
         particular with regard to the CCTS modules

       - We will continue work on a minor revision strategy,
         recognizing that we already have a workable proposed
         solution that we nevertheless need to discuss further

      Remaining issues for next round of NDR team work in Atlantic
      calls:

       - Whether namespaces change in minor versions

       - If so, how namespace prefixes are to be versioned

       - Whether we need to use substitution groups to effect
         minor versioning

       - What degree of backwards incompatibility we can tolerate

       - How today's decision affects NDR changes decided
         previously

       - Issues that still remain on the issues list

      AGREED: The Atlantic TC meetings of 5/18 and 5/25 will be
      devoted to deciding as many of these NDR issues as possible.

      AGREED: StephenG will wait on creating the list of schema
      changes for input to GEFEG until the beginning of June.

      TonyC: Members of the ver team should continue prototyping
      so that we can be crystal clear on the proposed solution.
      Need to look at counterexamples from MarkC.  There may need
      to be content changes.  [Didn't get this clearly -- Jon]

SUMMER VACATIONS

   JonB: Reduced availability from beginning of July till our TC
   meeting in Ottawa.

   DavidK: Unavailable Second half of July.

   TonyC: Unavailable first three weeks of June.

   StephenG, MikeG, MavisC: Available all summer.

   PaulT: Unavailable first two weeks in July, last week of
   August, first week of September.

   JonB: We need to check on MarkC's schedule, but it looks like
   we can keep working on NDR issues over the summer even if I'm
   not able to chair.

NOTE THAT THERE ARE NO MORE ATLANTIC TC CALLS TILL 18 MAY 2005.
However, there will be several Pacific TC calls during that time,
and work is expected to continue uninterrupted in the HISC and
SBSC.

Jon