Minutes of Atlantic UBL TC call 1 March 2006

[jon.bosak@sun.com]

MINUTES OF ATLANTIC UBL TC MEETING
16H00 - 18H00 UTC WEDNESDAY 1 MARCH 2006

ATTENDANCE

   Jon Bosak (chair)
   Peter Borresen
   Mikkel Brun
   Tony Coates
   Betty Harvey
   Bryan Rassmussen
   Zarella Rendon
   Sylvia Webb

STANDING ITEMS

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

      None.

   Subcommittee and team updates to Pacific call status reports

      None.

   Review of Pacific call

      No comments.

   Schedule review
      http://lists.oasis-open.org/archives/ubl/200602/msg00107.html

      (Regarding new line items 5 and 13 in supp0228)

      JB: (Line 5) What code list additions need to be made to
      NDR?

      TonyC: MartyB made some changes to the code list schemas,
      and these may not have been reflected in the NDR.  There
      probably is an outstanding item here.

      ACTION: JB to contact MartyB regarding possible code list
      contributions to the NDR.

      JB: (Line 13) It's been decided to copy the genericode
      schema into the support package for code lists (TonyC:
      Yes). What further needs to be done to gc 0.3?

      TonyC: Added metadata and a change to allow empty code lists
      having only metadata.  Will deliver the week of 3 April.

ACTION ITEM REVIEW

   ACTION: MB to verify that dk can rent a conference room Thu 25
   May.

      PB: The Northern European UBL group will fund this.  We can
      assume that it's being taken care of.

      ACTION: PB to discuss room logistics for 26 May with MavisC.

   ACTION: JB to check on possible Sun office in Brussels.

      JB: In progress, but the office is not in Brussels but
      rather Zaventem, and it's probable that they too will be
      closed on 26 May.

      AGREED to hold the work session 26 May at a rented room in
      Brussels rather than trying for Zaventem.

   ACTION: JB to ask Mark Leitch to attend the TC call 3/29 and to
   verify that he will be at the meeting in Brussels.

   ML (participating briefly): Funding not clear.

   ACTION: JB to contact funding sources in UK.

   SW: Will BH be in Brussels? (BH: No.)

      AGREED that SW will input the revised spreadsheets in
      Brussels to ensure correctness, then BH will generate the
      revised schemas the week of 29 May.  (We note that SW will
      be on vacation that week.)

   ACTION: BR to send us an updated proposal including a more
   extensive use of schematron. TC to review and
   be prepared to discuss in the Atlantic meeting 1 March 2006.

      BR: Did send the revised proposal:

      http://lists.oasis-open.org/archives/ubl/200602/msg00099.html

      The group discussed the NVDL possibility that was raised in
      this week's Pacific call and followed up in email to the ubl
      list from GKH and BR.  If we grant the requirement to allow
      ad hoc extensions identified by dk, then we have two
      alternatives: XSD ANY (as proposed by BR) and NVDL (as
      proposed by GKH).  The group was polled informally with the
      following results.

       - XSD ANY (the current proposal): strict validation will
         force a search for a schema associated with the namespace
         of the ANY section, but this presents a problem if the
         schema corresponding to that namespace is not available
         from a central location. Furthermore, assigning a
         namespace to the ANY section prevents reuse of elements
         in the parent schema.  If the namespace is omitted, it
         becomes difficult to distinguish items in an instance
         that derive from the main schema and items allowed by
         ANY.  We are generally agreed that ANY is a hack.

       - NVDL: exists because of the weakness in xsd; technically
         is probably the best solution.  The big worry here is
         about the availability of implementations and the lack of
         NVDL experience among implementers.

       - In any case, we will have to specify processing rules for
         extensions (e.g., if you don't recognize the namespace,
         go down to the next level; or, if you don't recognize the
         namespace, skip that entire subtree).  Given that such a
         rule has been established and you want to deal with an
         extension, ANY is a processing step, whereas NVDL is a
         preprocessing step.

      ACTION: TC to study BR's revised proposal (see above) and
      comment on the proposal; TC to think about NVDL and share
      comments with the group.

OTHER BUSINESS

   JB: Note that there will be no Pacific TC calls for the next
   three weeks.  There will be Atlantic TC calls for all three
   weeks, but they with be formally considered "work sessions"
   rather the TC meetings.

Jon Bosak
Chair, OASIS UBL TC