W3C XML Protocol Working Group teleconference, 10 April 2002

1. Roll call.

Present 27/24 Excused Regrets Absent

2. Agenda review, AOB

-- DavidF: we may cancel May 1 telcon because Europeans (except for Brits)
have a holiday. I'll come back on email about this.

-- MarcH: I'd like confirmation for what I proposed in my email:
http://lists.w3.org/Archives/Member/w3c-xml-protocol-wg/2002Apr/0086.html
WG agrees to proposal (no objections).

3. Approval of last week's (3 April 2002) telcon minutes: No objections

(one correction by Nilo) and so the posted minutes are approved.

4. Review of Action Items

5. Status reports:

Primer:
Nilo: incorporated editorials, and a few other changes based on
xmlp-comments and resolutions of issues.
DavidF: Is the Primer in sync with the specs which are being updated on
something like a daily basis.
Nilo: I'm tracking xmlp-comments and incorporating the changes that
are posted there, and by this mechanism I believe I am able to keep
the Primer in synch.

Spec:
Gudge: we are working on both parts, incorporating comments, in fact we're
looking for more work.
Other editors: Gudge is looking for more work, that is.
Marc: editors' todo list is pretty empty.
DavidF: I received concerns about readability of the spec so I am calling
for a few people (incl. me) to work on the spec over the next 2-3 weeks and
to propose changes to improve readability - strictly grammar and readability.
Any volunteers?
Colleen, Nilo, Highland volunteered.
Henrik: when will this activity start?
David: we'll work that out in the next day or two.

TBTF: didn't meet this week, meeting tomorrow, not much to report at the moment.

Conformance:
Anish: working on synching up with the Mar 23 spec snapshot. All of us 
(conformance team) are done with our respective sections. Oisin has been integrating the
changes we proposed, but he's been sick for the last week or two. We
hope to have an update of the document by the end of the week.
DavidF: could more volunteers help you?
Anish: once we have the update, we can paralelize more easily.
LynneT volunteers, she will contact the conformance team.

Usage Scenarios:
JohnI: changes in response to Yves, latest version is on the website now.

Requirements:
Looking into clarifying usage of the term block.
Henrik: it might have been done by removing the term.

Email binding:
Highland: I've sent to dist-app my plans on what I'm going to do.
I'll have that released by the end of this week.

6. Review of Last Call timetable

DavidF: we need closure on our issues list and set a deadline (EOB
April 11th), after which the new issues will be considered last call
issues. We will have to start up a last call issues list. There are
effectively already some issues on that list.
No objections from WG for this proposal.
DavidF: we also need a target (this call and the next two) for
finishing the issues.
DavidF: after that third telcon, the editors should prepare another
"checkpoint" document. After that, we'll also want to get the
"checkpoint" versions of all the other docs lined up for last call.
Editors, how much time will that check-point take?
Gudge: shouldn't take long. We'll need to change the namespace names
and the status sections.
Anish: this can take us (conformance team) a week and a half more. I
don't think we need more people than we have now - five.
DavidF: note that the Last Call documents are cimply working drafts,
nothing special. We'll have a cover letter where we specifically ask
for feedback where we feel we need it.
Henrik: surely conformance doc has to be ready at same time as
snapshot?
Anish: Are you asking whether the conformance document will be in synch
with a snapshot that is produced shortly after the 24th?
Henrik: yes: A spec snapshot on 24th, and a conformance doc snapshot on 24th.
Anish: if spec changes are relatively small, conformance doc should be
in sync, If the spec changes are larger, it could be difficult to keep the docs
in sync
DavidF: We understand the issue. Let's _plan_ for all the docs to be synched
and snapshots available on or around 27th.  All docs snapshot at same
time.  If changes in the spec greatly affect conformance doc and potentially
will cause slippage, the conformance team to raise a red flag. We have to
assume for now that the conformance doc will be in sync.

DavidF: how long after the current estimate (Apr 29th) will the WG
need to review these documents?
Noah: the readability review may increase the time we'll need
because some changes might be seen as changing the meaning of the
spec.
Noah: of course this depends on the volume of changes.
Henrik: major editorial changes can go into the Last Call issues
list. We don't want two major reviews.
Stuart: what are the dates for the review period? Early in May is the
WWW2002 conference.
DavidF: we'd have a week between the date and the start of WWW2002
conference.
DavidF: after the review we'll have things to fix but I think they
should be fairly minor things.
Noah: we run into the risk of having two major, Last Call reviews. I
think it does need the attention, but I'm nervous about the schedule.
Henrik: me too.
DavidF: I had not imagined the readability review to result in major
rewrites. Therefore I did not think it would need a second major
review.
DavidF: if we have the checkpoints on 29 April, we'll have a week to
review that, and the readability people will discuss what we'll do.
JohnI: if the readability review starts pretty soon, the reviewers
should have a good feel of how big the readability changes will be.
Noah: I've just reread the whole thing and it's a mix - some sections
are nicely written, some isolated sections should be rewritten, and in
many places it's just words here and there.
DavidF: so the readability folks will meet and discuss this, for the
time being we'll proceed with the schedule above, the readability
folks will signal us what they think.
DavidF: On May 8 (or 15 depending on WWW2002 impact) we can have the
go/no-go telcon for Last Call.

7. Identifying loose ends

DavidF: are there issues in email, that have not gotten into the list? 
No-one raises any issues.

8. Five issues to take as one block,

DavidF: can we decide on all the issues in one decision? All issues
54, 29, 56, 188, 36 - their proposals are on the agenda.
Anish: question about 36: The proposal contains wording that
applies to assertions but not to the tests. There can be an engine
that conforms to the assertions that does not pass these concrete
tests.
Noah: this is related to the fact that SOAP does not specify a processor.
DavidF: the motivation for the conformance suite is to be an aid for
understanding of SOAP and for helping to build interoperable implemantations.
Noah: we'll need to be very careful about how we say this.
Anish: I poropose we update the conformance document's introduction and
bring a new introduction back to the WG for approval at next week's telcon.
DavidF: is this acceptable to the WG?
No objections.
DavidF: OK, we will consider issue 36 next week. Can the WG close the
remaining issues (54, 29, 56, 188) with the proposals listed in the agenda?
No objections.
DavidF: those issues are closed.

9. New issues

-- 199: xml:lang on faultString
Marc: should we allow multiple faultStrings?
Gudge: my gut feeling is "put it in detail"
Gudge: I'll write up a proposal for next week that allows xml:lang on one faultString.

-- 200: SOAPAction header, action parameter on media type
Henrik: this poses an interesting question about dependency between our spec and the
media type draft. I will write up a proposal for next week: we'll live with the
dependency.

-- 201: inconsistency in the spec regarding structure of the body
Marc: in one part of the spec we say we mandate *nothing* about the
structure of the body, in other part we introduce Fault which does
mandate the structure.
Henrik: we had a proposal as part of proposal for 192. That might impact this issue.
DavidF: we'll postpone 201 until after 192

-- 202: definition of intermediaries
Herve: definition of SOAP intermediaries deals only with cases where
messages are forwarded becase a block requested it or the MEP
requested it. I think we need to take into account other causes for
intermediaries.
Henrik: the spec says that you do forwarding in reaction to header
blocks, or actively. I think it's already there.
Herve: it's in part there, a small modification will improve the
understanding of this part of the spec.
DavidF: take this to email, and we'll deal with it next week.

10. Issues

-- 192: when is a Fault a Fault?
A proposal appeared today so we might have had too little time to see it.
Stuart: we might want to get Mark Baker's attention to this.
Henrik: Mark was sitting beside me when I was composing this and he
seemed OK with this.
DavidF: it appears not enough people have read this, so we'll take it to email and
discuss it next week.

-- 195: mandating local name and namespace name
Jacek summarized briefly the proposal.
Noah and Henrik expressed some concern about losing the term return value.
DavidF: we'll get back to this, we should check if affects in any way what the XMI
folks have asked about regarding using the SOAP encoding for XMI.

Meeting ended.