W3C XML Protocol Working Group teleconference, 11 June 2003

1. Roll

Present 16/12 Excused Regrets Absent

2. Agenda Review, AOB

The chair reminded the working group that testimonials are due soon.

3. Approval of June 4 telcon minutes,

http://www.w3.org/2000/xp/Group/3/05/28-minutes.html

No discussion. No objection to approval of the minutes.

4. Review action items, see

[http://www.w3.org/2000/xp/Group/Admin/#pending]

to Chair - publish usage scenarios - pending
to MarkN, Noah, Herve - provide description - done
to Jeff - post notes from wsdl 1.1 binding for soap 1.2 - done
WG - review application for media type - done
MarkN - submit media type - pending
Jacek - send closing text for 429 - done (no pushback)
Noah - write up PASWA/XQuery data model - done

5. Status Report


* Planning for next f2f meeting (July 24 and 25, Sophia Antipolis,
France).

Staff have published a Web page with hotel details. Agenda TBD.

* Status Report - Creation of "Q+A" regarding SOAP 1.2 in a WSDL 1.1
environment How to proceed with the Q+A. To-date, we have had some very
limited
discussion (
http://lists.w3.org/Archives/Member/w3c-xml-protocol-wg/2003May/0058.html),
and we are awaiting publication (JeffS on dist-app) of notes regarding the
SOAPBuilders SOAP 1.2 binding for WSDL 1.1. If we are going to have
anything prepared in time for Recommendation (possibly June 24th), we
should draft it this week so it can be reviewed at next week's meeting in
time for the following week's Rec announce.

MarkN - would we be echoing the WSDL WG's temporary recommendation to use
SOAPBuilders?
Chair - that's one possibility.
Anish - there was discussion about forming a task force for doing more
work on the SOAPBuilders work. Have we reached a decision on that?
Chair - No. We only have 13 days.
MarkN - Do we have a sense of when WSDL will finish?
Chair - Sense is end of 2003 for LC. WSDL WG's next big task is SOAP 1.2
binding. Primary motivation today should be regarding what we should say
re: rec.
Jacek - I agree with the WSDL WG's recommendation. SOAPBuilders binding is
good enough. It may have problems, but it's realistic for now.
MarkN - +1
Jean-Jacques - +1
Chair - I'd like someone to take the text that have floated around and put
together a couple of paragraphs which answer the question. 1) SOAP has no
direct dependencies on WSDL 2) we recognize that people write to WSDL
rather than SOAP, and 3) there is a partial interim solution, which is the
SOAPBuilders binding.
Noah - Is this something that we expect to put in the press release? What
are we trying to draft here?
Chair - I think that what we're aiming to do is to create a Q+A that we
wouldn't proactively put out as part of press releases.
Noah - We don't have to say that most people are doing SOAP through WSDL;
we can say "many." Also, we can highlight that there's a WG responsible
for WSDL, and that we expect that the formal answer will result.
Chair - Will take that as a friendly amendment.
MarkN - Do we need to hide this? Would like to see a statement on the home
page.
Chair - That was our previous understanding. If we get the question a lot,
that would be a good idea.
MarkN - Seems reasonable.
Chair - Doesn't seem to be any objection to the core position,
notwithstanding details of distribution.
Jean-Jacques - I can do it.
Chair - Can you get a draft done by EOB (your time) Monday?
Jean-Jacques - Yes.

* Status Report - Registration of "application/soap+xml",
http://www.w3.org/2002/06/registering-mediatype.html. See update and
thread starting at
http://lists.w3.org/Archives/Public/xml-dist-app/2003May/0148.html.

MarkN - in progress.

6. PR issues, see http://www.w3.org/2000/xp/Group/xmlp-pr-issues.html

PR issues will be resolved and trickled into an Ed Copy in the same manner
as CR and LC issues except that (a) the bar for changes is higher, and (b)
the final Ed Copy will be 'submitted' to the Director who may choose to
accept/reject any of the changes for inclusion in any Recommendation
document(s).

* Issue 433
Tony - Reading the primer, the paragraph following the example of the
header intimates that if there are several mandatory header blocks which
fail, all of their QNames will be returned in the fault, which is in
conflict with section one. Suggestion is to change 'would' to 'could'.
Anon: +1
Chair - Proposal is to change 'would' to 'could' in the primer sentence
identified in the issue. No discussion. No objection. 433 closed.

* issues resulting from AC review
Chair - Staff, are there any AC issues to deal with?
Yves - Nothing that I am aware of.
Chair - If anything comes up, it will be resolved on the 18th of June.
That is the last meeting before we're supposed to go to REC. Are you aware
of that?
Yves - [no answer; chair assumes awareness]

7. Attachments


* Update on Copyright and IP statements.

Skipped; no pressing questions.

* Attachment requirements, see
http://lists.w3.org/Archives/Member/w3c-xml-protocol-wg/2003Mar/0027.html.

Skipped; no pressing questions.

* MarkN, Noah, Hervé report on abstract feature and conrete
implementation descriptions

Chair - general response from the group?
Jacek - I generally like it, I explicitly like the idea of renaming from
inclusion to optimisation
Gudge- generally, it looks reasonable (given its newness). Couple of
specific points - re: types discussion; I'm uninterested in optimising
floats and integers; raw binary is my interest.
Noah - clarifying my concern - the real question is if we only do base64,
nodes know that there is base64 typing; does the receiver know the
properties? Or do we formally model that the received envelope includes
the property?
Jacek - it also depends on how we model the sender side; if we model the
hint in the properties, I would expect it to be carried outside the
envelope.
Noah - the proposal as drafted has properties as a sender.
Jacek - then I expect that the receiving binding will expose these
properties.
Noah - currently, it's drafted so that they're not exposed above the
binding.
Gudge - I think I'm with Jack; it's an implementation detail as to whether
it's available to the application.
Noah - do we want to provide relay rules for it?
Gudge - This is somewhat related to MarcH's issue.
Jacek - it's an issue of whether the contract is formal on the receiver
side; it already is on the sender side.
Gudge - I don't think it matters if the same nodes are optimised in each
hop.
Noah - So, you want the contract to be hop-by-hop, not end-to-end.
MarkN - I see it as a hint, not a contract.
Jacek - I think the relay rules could be a SHOULD, allowing change.
Noah - there are a number of policy decisions; should an incorrect hint
generate an error on the sending side?
MarkN - is the property surfaced in the Infoset?
Noah - mixed feedback
Jacek - I imagine that some code must set the property to some value. I
don't see why the binding would not be able to set or change the value of
that property. When the feature gets implemented, the property would be
set, wherver it was set, to a set of nodes, and those would be optimized.
The implementation would layer it as a hint, and we could specify it as an
instruction.

* 430, semantics of attachments and intermediaries, Q1 in
http://lists.w3.org/Archives/Public/xml-dist-app/2003Jun/0002.html

[ incorporated into discussion above ]

* 431, how does binding determine what to serialise as attachment, Q2 in
http://lists.w3.org/Archives/Public/xml-dist-app/2003Jun/0002.html

[ incorporated into discussion above ]

* PASWA and XQuery datamodel (e.g.
http://lists.w3.org/Archives/Member/w3c-archive/2003Feb/att-0158/data-model.html
)

Noah - the two varients are those that do and don't transmit type
information. It seems to me that in paswa we're doing something very close
to the query model, and we should consider the synergies. I.e., "can we
send a SOAP message and get a query result?" If we connect the two, the
discussions about data models, signatures, etc. becomes easier.
Chair - reactions?
Gudge - I've read it, but I don't have a formal opinion as of yet.
MarkJ - I think it's interesting. The more that synergies exist, the
better. I fear complexity.
Chair - This will be taken to e-mail. Will continue discussion until the
WG determines whether there is synergy.

8. Recommendation-related activities

Janet - Testimonials call is out; open until 21 June. I have heard from
some
WG members as well as people outside the WG that want to make supporting
statements. I'd like to see more, esp. from inside the WG! I'll be happy
to help coordinate. There has been a lot of coverage of SOAP 1.2; for the
most part it has gone well. Two articles out of 23-24 that misunderstood
the work's status; most understood that the work was PR. I anticipate
another run of interest in SOAP1.2, because it will be released before
most press will be on vacation. Testimonials will help, so please mention
support, product support, etc.
Janet - Re the FAQ - will help in identifying compelling issues for
customers and developers. I rely on the WG to come up with issues from
customers, dist-app, etc. that highlight what SOAP 1.2 delivers. Please
send them into the team list for the FAQ sheet; I hope to have a draft
before your next meeting. Questions?
Chair - We've been discussing the Q&A regarding SOAP 1.2 in WSDL 1.1. One
of the aspects was whether we'd surface it.
Janet - Do we have a sense of if there are orgs that are holding off on
product announcements?
Chair - Are we still on track for a REC announcement on the 24th?
Janet - Yes. The new Webmaster is in place, so it should go smoothly.

Meeting adjourned.