XMLP WG telcon minutes, 8 September 2004

1. Roll call. Scribe for minutes selected from attached list. Actions to be
recorded on IRC.

Present 13/10
BEA Systems, Mark Nottingham
Canon, Herve Ruellan
IBM, Noah Mendelsohn
IBM, David Fallside
IBM, John Ibbotson
IONA Technologies, Suresh Kodichath
Microsoft Corporation, Martin Gudgin
Oracle, Anish Karmarkar
SAP AG, Gerd Hoelzing
SeeBeyond, Pete Wenzel (scribe)
Sun Microsystems, Marc Hadley
Sun Microsystems, Tony Graham
W3C, Yves Lafon

BEA Systems, David Orchard
Canon, Jean-Jacques Moreau
Microsoft Corporation, Jeff Schlimmer
Oracle, Jeff Mischkinsky
SAP AG, Volker Wiechers

Nokia, Michael Mahan

2. Agenda review, and AOB
The Important Stuff: CR implementation testing and CR issue resolutions

3. Approval of minutes, 11 Aug (840 + 5)

No modifications requested; approved without objection.

July 28 minutes have been posted, but not yet approved.

4. Review action items, see http://www.w3.org/2000/xp/Group/Admin/#pending.
These action items are taken from the XMLP member page.
 2004/08/04: MarcH
       Review WSDL Part 3, Due 2004-09-06
       WG to review and decide whether or not to send these comments to WSD

MarcH: Nothing critical.  Introduces a default binding, but does not
extend it to interface faults.  Would be easy and useful to do that.
Relationship between module and feature needs to be cleared up.
Other areas lack clarity.  HTTP binding did not capture some small items
as well.

DavidF: Any disagreement with this commentary?  No, so we will submit
to WSD WG.  MarcH will forward his email to them.

 2004/08/04: MarkN
       Review WSDL Part 1, Due 2004-09-06

Change due date to Sept 22.

 2004/08/04: MikeM
       Review WSDL Part2, Due 2004-09-06

Change due date to Sept 22.

 2004/08/11: Editors
       Send email regarding change to section 3.2 of MTOM to dist-app

 2004/08/11: Editors
       Change all namespaces from /YYYY/MM/ to /2004/08/

 2004/08/11: Yves
       Arrange a concall on Zakim for MTOM implementers at 11am PST
       See agenda item #6

5. Status reports and misc (9.00 + 10)
-- Media types registrations, i.e. "application/soap+xml" and
"application/soap_xop+xml" (MarkN)

MarkN: It has progressed to "rfc-editor" stage.  One more step + 48
hours before it is finally published.  Estimate 1-2 weeks to get there.
Request W3C Team to review for editorial nits, such as references.
Change controller should be updated to the appropriate W3C contact:

MarkN will verify that this change can be made without reverting to a
previous state.

Yves: application/xop+xml registration sent to IETF.

-- XMLP/WSD Task Force and WSDL Media Type document (Anish)

Anish: Nothing to report.

-- f2f meeting

DavidF: CR period ends Sept 15.  We have received 5 comments.  Is it
still necessary to meet?

Noah: Any issues about the WG's future can be discussed by phone.

DavidF: Purpose of this meeting is to get us to PR.  Unless we get
comments that send us back to pre-LC, doesn't seem like we will need to
meet F2F again.  Leave it pencilled in; will make the decision next week.

Noah: Will miss next week's call, but could attend F2F.  Feel free to
make any decision about whether or not to hold it.

MarkN: No problem cancelling hosting arrangements, but personally have
already bought ticket.

DavidF: Nilo is expecting to attend; will email him a warning that it
may be cancelled.

-- Primer review,
http://lists.w3.org/Archives/Public/xml-dist-app/2004Jul/0006.html, WG
members are asked to review and post comments

6. Candidate Recommendation (9.10 + 50)
-- Status of testing: what end-points are available, did the MTOM
implementers telcon take place and if so what transpired, have any messages
been exchanged, what plans for testing, is there a complete set of tests
for all parts of the spec?

Gudge: Attended the call with Jean-Jacques.  Agreed that the proposed
tests were ok.  Are being used this week between Canon, Microsoft,
White Mesa.  All positive tests are working.  Microsoft<->Canon,
Microsoft<->White Mesa.  Failure cases will be attempted later this week.

Only issue is that "start-info" parameter was spelled without hyphen,
because SOAP example incorrectly excludes it.  That has been corrected
in the implementations.

Herve: Canon has completed all 7 test cases against both Microsoft and
White Mesa.

Anish: Should complete writing tests for Resource Representation Header
over the weekend.

Gudge: Implementation can handle resource header, but there is no
application-level code that actually uses the data.

DavidF: Do we need to retest to capture logs?

Gudge: Have complete logs of everything sent & received.

DavidF: Send me a complete set of logs; will post on a web page.

Gudge: Will have Bob send an email regarding HTTP headers.

-- Issues
     o 500, XOP, (editorial) broken references section,

DavidF: Propose that this is editorial, and assign to editors for
resolution and closing email.

Yves: References section of all 3 documents need to be updated.

Gudge: Links disappeared when CR versions were generated.

     o 504, Rep, (editorial) i18n issues,

Editors to fix and send closing email.

     o 501, Rep, i18n issues - encoding and language,

Noah: We tunnel data in a base64 octet stream, and add metadata.

MarcH: There is an issue with inheritance of charset.

Gudge: To Andrea's point #3, the SOAP document is in a single charset.
Our answers to 1-3 would be:
Yes, base64 is required.  Yes, after decoding the base64 element,
its contents may be in a different charset.  This is not a problem.
Charset of decoded data may be determined in any number of ways.

Gudge will draft a response, send to dist-app.

What about point #4?

MarcH: We allow attribute extensibility.

Yves: Is not appropriate to use xml:lang.

Gudge: base64 doesn't have a human-readable language.  Someone else
could write a specification to define how to do this.

DavidF: Or use an existing mechanism.

Gudge will address this in the proposed response as well.

     o 502, Rep, i18n issues - URI handling,

Noah: We assumed they were IRIs, but Schema says they are URI.
The reference is HLink spec, rather than something more normative.
We didn't mean to use IRIs here; they are not visible on the web.

Noah: Change our spec to read that the attribute value "SHOULD be a URI".

MarcH: Why not "MUST"?

DavidF polls: Group preference is for "SHOULD".

Noah will draft a proposal answering question #5.

DavidF: We are over time; meeting is adjourned.

     o 503, Rep, i18n issues - HTTP semantic,

 7. SOAP 1.2 Recommendation maintenance (postponed)
 -- Issues and proposed resolutions
     o Issue Rec25, http://www.w3.org/2000/xp/Group/xmlp-rec-issues.html#x25
 , "media type registration". Issue text describes a couple of ways to
 proceed. This issue is pending IANA registration of media type.