minutes from XML processing model WG call 2006-03-30

See: http://www.w3.org/XML/XProc/2006/03/30-agenda.html

XProc Telcon 30 Mar 2006

The XML Processing Model (XProc) WG met on Thursday, 30 Mar 2006 at
11:00a EST[1] (08:00a PST, 16:00GMT, 18:00CET, 01:00JST+, 09:30p
India) for one hour on the W3C Zakim Bridge.


  Andrew Fang, PTC-Arbortext
  Paul Grosso, PTC-Arbortext
  Rui Lopes, (public) Invited expert
  Murray Maloney, W3C Invited Experts
  Alex Milowski, W3C Invited Experts
  Michael Sperberg-McQueen, W3C/MIT (chair pro tempore in Norm's
  Richard Tobin, University of Edinburgh
  Alessandro Vernet, Orbeon, Inc.
  Mohamed Zergaoui, INNOVIMAX

Regrets or otherwise absent:
  Robin Berjon, Expway
  Erik Bruchez, Orbeon, Inc.
  Vikas Deolaliker, Sonoa Systems, Inc.
  Jeni Tennison, W3C Invited Experts
  Henry Thompson, W3C/ERCIM and Univ. of Edinburgh
  Norman Walsh, Sun Microsystems, Inc.

 1. Administrivia
      1. Regrets: Henry, Norm; Michael to chair.

Also regrets from Erik Bruchez.

      2. Accept this agenda[3].


      3. Accept minutes of 23 Mar 2006[4].

Accepted without change.

      4. Next meeting: 6 Apr 2006.

Regrets: Richard Tobin (6 and 13 April), MSM (6 April).

      5. Face-to-face meeting planning.

[At this point, Alex Milowski and Mohamed Zergaoui joined the call.]

MM said he had heard back from some people, but not from MSM or Paul;
they both said they could make either set of dates.  MM said two
people have expressed a preference for the earlier dates.

So the question is now: shall we have such a ftf meeting?

PG proposed that we agree on the earlier dates and then ask how many
would attend; have the meeting if there will be critical mass,
otherwise don't.

Question: who will attend if the WG has a face to face meeting 2-4
August near Toronto?

  Yes: Michael, Murray, Alex Milowski, Mohamed (probably)
  No: Andrew (probably)
  Don't know:  Alessandro (and Erik), Richard (unlikely, though

All don't-know answers reflected budgetary uncertainty.

We left open whether to have the ftf meeting or not.

 2. Technical
      1. XProc Requirements and Use Cases[5]

AM said the one thing we need to change is to add language saying we
aren't actually committing to solve all of these use cases in 1.0.

We discussed first the general question: do we want such a change?

RT argued that the name 'requirements document' will otherwise give
readers the wrong idea.  We should be explicit about what we really

AV asked whether there is any specific use case people have in mind
when they say we're not committed to supporting it?  Perhaps we should
remove those?

AM gave the DSDL use case as an instance. He doesn't (he said) know
what it means.  If we can usefully support it, fine, but let's not
commit ourselves to it without more clarity.

RT said he thought that lots of the use cases described were unlikely
for 1.0.  E.g. anything that refers to processing non-XML data: yes,
it shows up in use cases, but providing a full description of how to
do that is just way too complicated to be low-hanging fruit for 1.0.

MSM noted that schema support doesn't seem to have consensus as a
required thing that people must support.  

[Scribe's note: no one actually mentioned the use cases which require
iteration to a fixed point, although those who were present at the
Cannes face to face were fairly explicit that we should not commit
ourselves a priori to support that form of iteration.]

MSM asked AV if he were content with the answers; AV said yes, OK,
there are also some use cases which he also feels uncomfortable about.

The sense of the group, MSM suggested, was that yes, we wanted wording
of the sort described by Alex.

Alex Milowski proposed:

  This section contains a set of use cases that support our
  requirements and will inform our design. While there is a want to
  address all the use cases in the specifications that are designed to
  meet the requirements in this document, in the end, the first
  version of those specifications may not solve all the following use
  cases. Those unsolved use cases may be address in future versions of
  that specification.

The WG felt that "that specification" at the end should probably be
"those specifications".

There was also some concern about the second sentence, which MSM found

After a little more discussion, we decided (in view of the time left
in the meeting and in view of our desire to publish the document on 4
April) to leave it to the editor.

RESOLVED: to instruct the editor to put in some wording with a warning
about the requirements.  The wording proposed in the meeting may be
taken as a basis; the editor should take what advice he can get from
others in the WG and then use his best judgement.

We then asked ourselves whether we wanted similar wording for the
requirements themselves.  RT noted that there is at least one
requirement (4.10) which RT is not happy with as a minimum requirement
for conformance.

AM argued against general wording about the requirements not being
requirements.  It IS, he noted, a requirements document.  It would be
better to add a note to the controversial one saying "This was doesn't
have consensus yet" than to put in a general note saying none of them
are firm.

[At this point, Murray Maloney left the call.]

MSM suggested adding a Note to 4.10 saying something like:

    Note: there is not consensus in the Working Group that full
    support for the XPath 2.0 Data Model should be required for
    minimally conforming implementations of these specs.

RESOLVED: to instruct the editor to use his best judgement on whether
to add a note to 4.10 noting that there is not consensus on it, and on
the wording of such a note.

ACTION: MSM to proofread the document.

We reconfirmed our earlier decision to publish the document:

RESOLVED: to publish the first public working draft on 4 April or

Alessandro asked whether we should try to include the new use case
that had come up in email; this is the one raised by Mohamed Zergaoui

MSM noted that since we need to give the Webmaster several days of
leeway to check our document before publication, if we plan to publish
the document next week we really need either to have final wording
today or give the editor (or some subgroup of the WG) authority to
agree on final wording on our behalf.  On the whole, he suggested, it
might be better to wait, discuss the new use case, and include it in
the next draft of the document.

So agreed.

Paul Grosso asked whether we should publish just the HTML of the
document, or also the XML?  The general sentiment seemed to be that it
might be better to publish the XML as well, but only if it were not
too much extra work.  

RESOLVED: to charge the chair, editor, and staff contact (i.e. Norm,
Alex, and Michael) with deciding whether to publish the XML source as
well as the HTML.

      2. Review of Richard's proposal[6]

We did a very brief and incomplete review of Jeni's comments on the

We'll take the topic up again next week.  RT noted that he will be
absent next week and the week after, and said he looks forward on his
return to seeing the metamorphosis of this proposal into someting
completely different.

      3. Conditionals and sub-pipelines

Not discussed for lack of time.

 3. Any other business


[2] http://www.w3.org/XML/Processing/
[3] http://www.w3.org/XML/XProc/2006/03/30-agenda.html
[4] http://www.w3.org/XML/XProc/2006/03/23-minutes.html
[5] http://www.w3.org/XML/XProc/docs/langreq.html
[6] http://www.cogsci.ed.ac.uk/~richard/pipeline.html

Received on Friday, 31 March 2006 00:41:27 UTC