[xslt-conformance] Minutes from August teleconference

[David Marston/Cambridge/IBM <david_marston@us.ibm.com>]

OASIS XSLT/XPath Conformance Voting Member Meeting

Minutes for August Teleconference 2002-08-28 13:30-15:30EDT

Roll call and quorum
   Present:
     Lofton Henderson
     David Marston, IBM Research
     Carmelo Montanez, NIST
     Tom Amiro, Sun Microsystems
     Jeff Kenton, Observer, Individual Member (2002-04-16)
     Dimitre Novatchev, Individual Member
     Zarella Rendon, Observer, Individual Member (2002-08-28)
   Regrets:
     (none)
   Absent:
     Pranav Kandula, Microsoft Corporation
     (and a couple observers we expected)

Agenda
  - amendments - discuss xsl:number
  - motion to accept: 1: Carmelo 2: Lofton

Minutes of the last meeting
  - review of minutes of last meeting 2002-07-25
  - motion to accept minutes of meeting: 1: Lofton 2: Carmelo

Membership
  - welcome to new voting members
    - Tom Amiro and Dimitre Novatchev became voting members at the end
      of the meeting. We now have six voting members.
  - welcome to new observer members
    - This is 2nd appearance for Jeff; he wants to join
    - This is the first appearance for Zarella; she stated that she
      wants to join 0n 2002-04-26
  - who will take on the role of Chair or Acting Chair?
    - Can ask again at September meeting
    - mail list change is required for chairs@lists.oasis-open.org

Reports of standing committees:
  - Comparing Output and Role of Canonical XML - no report
  - Discretionary Items (stable; no report expected) - no report
  - Web Site Maintenance - Carmelo will continue normal maintenance

Outstanding action items
  - 20000906-ACT-9: (Carmelo) model and prototype rendition control
    instance and XHTML production working with Ken's results
    - No contact between Dimitre and Carmelo as of meeting time
  - 20011217-ACT-1: (All) Test reviews, tracked by Carmelo
    - See web page for status
      http://www.oasis-open.org/committees/xslt/reviewedTests091102.html
      or follow link from "Work Page" as the date portion changes
    - Some Microsoft tests need a second review

Agenda items
  - How well does the test review system work?
    - Possible issue: reviewers want to give feedback about aspects
      of test other than pure conformance
    - How do we check for completeness of coverage? Deal with it later.
    - May want to improve or at least explain status codes as a workflow.
      To be discussed further in September.
  - Do we need to change anything about the submission process?
    - See note about status codes above
    - No observed need to change what is submitted
  - Do we need to change anything about the review process?
    - Probably don't need detailed (down to sentence) citations
    - Idea from Tom: make checklist from Review Policy
  - How close are we to issuing a pilot test package?
    - Referring to the infamous diagram at

http://www.oasis-open.org/committees/xslt/contributions/deliverables-20020425.gif
      we need the XSLT transform in the lower right, the stylesheet that
      is an input to it, and the "rendition control instance" XML data
      that specifies which discretionary choices apply to the particular
      processor under test. (Note that xsl:number has a few such choices,
      so we need this in time for the pilot release.) A questionnaire for
      developers exists, but the format of the answers can be adapted to
      the needs of this transformation. Probably also need to review the
      proposed "content" on the upper right, the Test Lab Guidelines, and
      the state of output comparison. For more about discretionary choices,
      see http://www.oasis-open.org/committees/xslt/xsltdisc.htm
    - Also see discussion below about reviews of certain xsl:number tests
  - How close are we to opening our call for submissions to more
    categories of test cases?
      - Discuss in September and vote on some
      - Those present in August seemed to favor "structural" areas like
        keys, axes in XPath, generated elements and LREs, copying, sort,
        xsl:apply-imports, generated attributes. Namespace nodes are seen
        as important to get into, but somewhat of a trouble spot. Jeff
        expressed an interest in the format-number() function.
  - Discussed some difficulties in xsl:number, but didn't reach any
    conclusions because not all interested parties were present.
    - Regarding Erratum E24 in
      http://www.w3.org/1999/11/REC-xslt-19991116-errata/
      there still seem to be vague aspects, like zero vs. null string
    - What to do for <xsl:number value="foo" format="(1)"/>? Most with an
      opinion thought "(NaN)", some thought "NaN". This of course assumes
      the discretionary choice number-not-positive=pass-through was made.
    - Committee expertise issue: how do we check language-dependent
      numbering for languages that none of us knows?
    - ACTION: David to send out separate emails for each known issue
  - Upcoming meeting dates
    - September: teleconference 25-Sept (Wednesday) at 13:30 EDT
    - October: tentative designation of 23-Oct, teleconference
    - No face-to-face happened in August. Should have one soonish.

New and old action items for next meeting
  - 20000906-ACT-9: (Carmelo) model and prototype rendition control
instance and XHTML production working with Ken's results
  - 20011217-ACT-1: (All) Test reviews, tracked by Carmelo

Next meeting dates
  - 2002-09-25 - 13:30-15:30EDT - teleconference monthly meeting
               - NIST to look into facilities

Acknowledgements
  - many thanks to NIST for provision of the teleconference facilities
    today

Adjournment
  - 14:55EDT