Minutes of Atlantic UBL TC call 31 August 2005

[jon.bosak@sun.com]

MINUTES OF ATLANTIC UBL TC MEETING
15H00 - 17H00 UTC WEDNESDAY 31 AUGUST 2005

ATTENDANCE

   Jon Bosak (chair)
   Mikkel Brun
   Marty Burns
   Tony Coates
   Mark Crawford
   Mike Grimley
   Betty Harvey
   Anne Hendry
   Sylvia Webb

STANDING ITEMS

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

   None.

Liaison reports

   None.

Subcommittee reports

   None.

Team reports

   Code Lists -- see below.

   Digital Signatures -- see below.

Review of Pacific and Europe/Asia calls

   No comments.

Schedule review:
   http://lists.oasis-open.org/archives/ubl/200508/msg00167.html

   NDR (row 8)

      JB: We are supposed to start work on minor versioning
      methodology.  This will be added to next week's agenda.
      "Minor versioning methodology" includes the question of
      whether to minor-version namespaces.

   "Finish implementation of 2.0 NDRs" (row 9)

      ACTION: SW to check on implementation of 2.0 NDRs per row 9
      of the schedule and report back on Monday.

   "Load phase 1 models" (row 10)

      SW: Have received models from PB and am working on a script
      to allow BH to load them into EF.

   "Validate schemas" (row 29)

      JB: I will need to be notified when the schemas are ready.

   "Define and document representative extended procurement
   scenarios" (row 30)

      ACTION: PB to begin work on defining and documenting
      representative extended procurement scenarios per row 30 of
      the schedule.

ACTION ITEM REVIEW

   ACTION: TC members visiting NYC in October or November to check
   out the Sun facilities and see whether they are suitable for a
   UBL TC meeting.  JB will be coming through in mid-November and
   will also check then.

      Pending.

   ACTION: MarkC to review schema import diagrams in the NDR
   document and update as appropriate.

      Pending.

   ACTION: JB to alert the TC that we are moving to closure on the
   NDR document.

      Done.

   ACTION: SW to give us a deadline for completion of a document
   describing our spreadsheet format to be inserted into the UBL
   2.0 schedule.

      Pending.

FOR THIS MEETING

   Proposed Catalogue work team

      AGREED to form the Catalogue work team as proposed (see
      Pacific minutes).

   NDR document revision approval

      MarkC: Haven't revised the diagrams yet, but the document is
      stable enough for tool configuration.

      AGREED that we will adopt the NDR draft posted 11 August as
      stable for purposes of schema generation.

   Phase 1 content model approval

      AGREED that we will accept the spreadsheets posted by PB 30
      August for initial phase 1 schema generation in order to
      begin the review cycle.

   Code list progress

      JB: We are moving ahead with the direction we adopted in
      Ottawa, but we need to flesh it out.  What is the impact on
      EF of the alternatives in which some CLs are just nmtoken?

      SW: Probably nothing significant enough to bring to your
      attention; will have an update Monday.  Class 2 CLs were not
      hardcoded; they were separate modules, so not referencing
      them should not be a major issue.

      JB: I'm guessing that there will need to be changes to the
      NDRs, but that we can make those changes without affecting
      the schedule.

      (General agreement that this is probably the case.)

      TonyC: Haven't done the translation to schematron yet, just
      the xsd case:

         http://lists.oasis-open.org/archives/ubl/200508/msg00184.html

      The script translates the new code form into 1.0 schema
      form, but haven't done all the metadata yet.

      JB: So this would provide a way to take the new instance
      form into the old XSD approach...

      TonyC: And also support the ACORD method.

      MartyB: GKH hasn't responded fully to the points raised in
      email.

      JB: So please continue that discussion.

      ACTION: MB to continue code list dialog with GKH.

      AGREED that redefine is not an appropriate technique for
      code lists and will not be considered further as a solution.

      MarkC: Continue to be opposed to this course.  We should go
      back to enumerations in schemas and not worry about
      customizations.  ATG is allowing union of CLs for
      customizers.

      ACTION: MarkC to send ATG paper on CLs to MB and share it
      with the list when that becomes possible.

      MarkC: Also continue to be opposed to use of substitution
      groups.  See xml-dev posting in last couple of days
      regarding SGs from HenryT on Tues.... Many reports of
      interop problems relating to choice groups in data-binding
      tools.

      MartyB: Without a mechanism like SBs, the recipient does not
      know what the sender intends to be the code list.

      TonyC: I want a choice in how I validate it.

      MartyB: Yes, but not reinterpret it.  Say I have my own
      version of EUR that has an inflation tracker built in.  With
      SGs you have to declare that.

      TonyC: But in reality I would not accept that unless you
      told me [out of band] what your substitutions actually are.
      I'm not going to leave my system open and just assume that
      what you've sent me is OK; I will want prior notice of the
      change.

      JB: You can always lie to me; substitution groups don't
      prevent that.  The point of referencing an ISO CL is to say
      that *this* is the meaning of EUR.  If you change the
      meaning of codes instead of just adding or subtracting
      codes, then you lose the whole point of standard code lists.

      SW: Agree that the meaning of codes is significant and can
      have legal implications, but do we need to have complex
      methodologies for interpreting codes?  Multiple validation
      passes and multiple methods of interpretation are too
      complex for SMEs.

      TonyC: We are trying to find ways to allow that as an option
      without making things too complicated.

      JB: We could look at this a different way: If we use TonyC's
      script to provide enum xsd versions of all the CLs, then we
      can say to the user, "We're giving you 1.0-style static
      validation, just as in UBL 1.0, and also this new two-pass
      method that you can use as an alternative."  Can we provide
      both versions of the CLs while using the same document
      schemas?

      (General agreement that this appears to be the case.)

      JB: So aside from the extra work of explanation and
      packaging, doesn't this satisfy both sets of requirements?

      MartyB: It still doesn't satisfy the requirement to
      guarantee the version of the code list in the instance.
      There's no audit trail.

      MarkC: Never got the documentation of CL version in the
      instance as a statutory requirement in DK.  UBL identified
      this issue for CEFACT, but we need to provide confirmation.

      MikkelB: Yes, it is a legal requirement that the CL version
      should be in the document instance.

      MarkC/SW: This was never a requirement in EDI.

      ACTION: MikkelB to provide a clear specification of what is
      required by law in DK regarding CLs by next week.

      (Returning to Marty's point)

      TonyC: This relates to customization; do they just edit the
      CL we send them and agree with their friends somehow, or do
      we want some more formal mechanism?

      JB: How can we prevent people from just changing the files?

      MartyB: We provide a form [substitution groups] that will
      encourage them to do the right thing.

      JB: We should continue to work on this over the coming week.
      If the two-pronged approach is possible, we should see
      whether we're willing to put in the work.

      TonyC: We need to define the schema format for CLs.  In
      particular, we need MartyC's input on the requirements.

      MartyC: We will need everything in the current document
      except abstract types and substitution groups.

      ACTION: MartyB to provide a revised example of the enum
      version of the currency CL as formerly proposed but without
      substitution groups, by next week if possible.

      ACTION: TonyC to provide a script that will transform a CL
      instance conforming to the proposed CL schema into a
      schematron schema by next week.

   We note that MikkelB, MG, and TonyC will not be on the Atlantic
   call next week, and that MarkC may not be able to attend as
   well.

Jon Bosak
Chair, OASIS UBL TC