Minutes of xmlprotocol WG telcon, 12 March 2003

Timestamps are in UTC.

1. Roll call
Present 17/14
AT&T	Mark	Jones
BEA Systems	Mark	Nottingham
BEA Systems	David	Orchard
Canon	Herve	Ruellan
DaimlerChrysler R. & Tech	Mario	Jeckle
IBM	David	Fallside (chair)
IBM	John	Ibbotson
Microsoft Corporation	Martin	Gudgin
Oracle	Anish	Karmarkar
Progress Software	Colleen	Evans
SAP AG	Volker	Wiechers
Software AG	Michael	Champion (scribe)
Sun Microsystems	Tony	Graham
Sun Microsystems	Marc	Hadley
Systinet (IDOOX)	Jacek	Kopecky
Unisys	Lynne	Thompson
W3C	Carine	Bournez

Excused
AT&T	Michah	Lerner
Canon	Jean-Jacques	Moreau
DaimlerChrysler R. & Tech	Andreas	Riegg
IBM	Noah	Mendelsohn
Microsoft Corp	Don	Box
Oracle	Jeff	Mischkinsky
Progress Software	David	Chappell
SAP AG	Gerd	Hoelzing
Software AG	Dietmar	Gaertner
Systinet (IDOOX)	Miroslav	Simek
Unisys	Nick	Smilonich
W3C	Yves	Lafon

Regrets
Ericsson	Nilo	Mitra
IONA Technologies	Oisin	Hurley
SeeBeyond	Pete	Wenzel

Absent
Fujitsu Limited	Masahiko	Narita
Fujitsu Limited	Kazunori	Iwasa
IONA Technologies	Eric	Newcomer
Macromedia	Glen	Daniels
Matsushita Electric	Ryuji	Inoue
Tibco	Don	Mullen

19:13:05 [mchampion] 2. Agenda review ...
19:13:41 [mchampion] Minutes from F2F not yet ready, therefore omit agenda item 3
19:16:25 [mchampion] Plan to end call after 90 minutes at latest
19:16:37 [mchampion] No additional items for agenda

3. F2f minutes review postponed until next week

19:16:47 [mchampion] Item #4 - Review action items 
19:17:39 [mchampion] Publish nln as note: Pending, perhaps by end of week
19:17:53 [mchampion] Gudge closed issue 415. Done
19:17:59 [mchampion] Editors: Done
19:18:09 [mchampion] Issue 416 done
19:18:53 [mchampion] Editors remove root attribute ... done
19:19:07 [mchampion] Editors remove ref attr .... done
19:19:27 [mchampion] Editors remove uniqueness constraint ... done
19:20:00 [mchampion] Editors fix schema to allow soap NS elements as immediate
children of header ... done
19:20:21 [mchampion] Chair draft message to WSD on use attribute ... done
19:21:00 [mchampion] Colleen Send comments on QA last call working draft ... done
19:21:59 [mchampion] But some pushback on deprecations ... 
19:22:33 [mchampion] DavidF Update feature 27.1 ... done
19:23:00 [mchampion] Gudge contact SOAP implementators on nodeType ... done
19:23:18 [mchampion] Gudge Propose text for section 2.7.4 in part 1 ... done

19:23:50 [mchampion] Item #5 Status reports
19:24:34 [mchampion] Part 0 - Nothing to report
19:25:08 [mchampion] Part 1 - All changes gone through at F2F have been made
19:25:40 [mchampion] Only change needed before PR is the final namespace NS URI
and boilerplate about "Proposed Rec"
19:26:01 [mchampion] Gudge: Need to check F2F minutes to confirm that all changes made
19:27:12 [mchampion] Jacek: Send mail about namespaces ... Will PR namespace be
same as Rec namespace?  Depends on date of publication, which implies that they
will be different between PR and Rec.
19:28:04 [mchampion] Chair: Should we push back on this?  Go to just using 2003
in NS URI? 
19:28:30 [mchampion] Gudge: We may have to make changes, and keeping month might
be better in long run.
19:30:24 [mchampion] ACTION: Editors and W3C staff should request that we can keep
the PR namespace if nothing changes, and to get an explanation if this is not possible
19:31:14 [mchampion] Editors: Nothing else on part 1 or 2
19:31:40 [mchampion] Question to editors: is diff version of spec available? No,
but they will produce one tomorrow
19:32:14 [mchampion] Test collection: Nothing to report, status same as at F2F
19:32:39 [mchampion] Registration of media type.  Yves not on call ...
19:32:53 [mchampion] nln publication - pending
19:32:59 [mchampion] IPR - nothing to report

19:33:16 [mchampion] agenda item 6. CR Progress and issues
19:33:35 [mchampion] DavidF: Nothing really to report.  Telcon Friday with implementers
19:34:46 [mchampion] Node Type: Gudge has three responses so far ... low on
everyone's priority list
19:35:53 [mchampion] Jacek also responded directly to list... people who don't
implement encoding don't need nodeType; 
19:36:01 [mchampion] Sometimes useful if no WSDL available.
19:36:46 [mchampion] BEA sez they will look into this with their implementers.
19:37:48 [mchampion] May also get answers from Apache and Paul Kulchenko .. good
to get responses from them
19:38:01 [Gudge] ACTION: Gudge to chase Apache, BEA and PaulK re: nodeType
19:38:03 [mchampion] before next week.  Then we'll make a decision.
19:38:34 [mchampion] DavidF: Operhaps 3 options: (i) Drop nodeType from spec,
(ii) go to PR with incomplete implementation evidence and say that we have minimal
implementations now and more possible in future.
19:38:49 [mchampion] (iii) try to persuade more people to implement remaining
features and check interop.
19:38:57 [mchampion] DavidF: Need more data before we can really decide.
19:39:18 [mchampion] ACTION: Gudge should ping BEA, APache, and Kulchenko about
nodeType implementation
19:40:10 [mchampion] Issue 417 .... Came in after we decided to go to PR net of
edits from F2F.
19:40:43 [mchampion] Chair: There's a high bar to changes in spec at this point. 
Not clear what changes are needed (no specific proposal was made), and we won't
consider anything
19:41:00 [mchampion] that doesn't include proposed text at this point.  
19:42:12 [mchampion] Commentator says - A new person wouldn't suspect that other
children are possible, so we can live with explaining it outside the spec.
19:42:58 [mchampion] Jacek (speaking as issue originator) Will not push for it
unless others care strongly, willing to withdraw issue
19:43:11 [mchampion] Item 6 - Pushback on closed issues
19:43:15 [mchampion] DavidF: none mentioned
19:43:47 [mchampion] ACTION: Issue list maintainer should close 417 with notation
that issue withdrawn 

19:44:37 [mchampion] Item 7  - Attachments
19:44:59 [mchampion] DavidF (reviewing agenda item): Considering two lines of moving
forward -- Concrete attachment scheme is one ...
19:45:53 [mchampion] would have met most of requirements.  Most favor concrete
implementation in short term.
19:46:02 [mchampion] Do we have consensus?  
19:46:32 [mchampion] Second part: Infoset based model of attachments from BEA and
MS.  Chair believes that many interested, but they are concerned about impact
19:46:59 [mchampion] on 1.2 spec and on details of what it would entail.  Extensive
discussion at F2F, see minutes.
19:47:40 [mchampion] Compared to first approach, more of a long-term thing.  How
can we gain a greater understanding of the second approach.
19:48:11 [mchampion] How do we  get a writeup of the second approach so we can decide?
19:49:14 [mchampion] DavidO: Action item to DavidF from F2F that's not on action
item list ...
19:51:03 [mchampion] Exact text may not have been recorded ...gist is  that the chair
should look into issue of how the WG can bring in technology from outside and what
the options are.
19:51:42 [mchampion] DavidO - Understood action for the chair to determine what
is in scope for WG vis a vis this specific technology, not the general procedural issue.
19:52:18 [mchampion] DavidF - no definitive answer, because it's not clear what
this "thing" is.
19:53:09 [mchampion] DO: The gist of the issue is that we don't want to waste
time on something that is ultimately out of scope.  Let's resolve the scope issue
in the next 1-2 weeks.
19:54:13 [mchampion] DavidF: One issue on scope is that the second approach appears
to rely on XInclude, and the use of XInclude requires changes to XInclude.
19:54:30 [mchampion] Multiple dependencies make a yes/no answer hard.
19:54:50 [mchampion] DaveO - The coordination issue is separate from the in-scope issue.
19:55:19 [mchampion] MarkJ: It could also be soap:include, not necessarily XInclude
19:55:22 [mchampion] Others concur.
19:57:53 [mchampion] DavidF: It would be very helpful for the proponents of the
2nd approach to make a more detailed description available.
19:58:16 [mchampion] DaveO: When would we need such a thing?
19:58:57 [mchampion] DavidF: ASAP.  Preliminary answer on scope possible in a week,
but still need more detail.
19:59:30 [mchampion] DaveO: Understands that there is a circular dependency, and
did promise a document by next week concall.
20:00:08 [mchampion] MarkJ - actually two weeks from F2F, so 2 concalls hence.  
20:00:23 [mchampion] DaveO: Will try before concall, but commitment was by Friday.
20:01:12 [mchampion] "It will have a URI" is the only promise vis a vis where it
will be published.
20:01:36 [mchampion] MarkJ - Can we move forward before then?
20:02:36 [mchampion] Will propose an ordering of requirements that aren't dependent
on the two actual proposals.  Wants to get issue settled ASAP, 
20:03:02 [mchampion] but want us to be productive while waiting for details on the
second approach.
20:03:46 [mchampion] ACTION: Mark Jones will prepare an index for the AF requirements
doc, in order of least to most dependent on the choice of the proposals under consideraiton
20:05:09 [mchampion] DavidF: Back to part 1 of agenda item 7, would like to get a sense
from people who were not at F2F whether they wish to see a concrete attachment
scheme as a work product of the WG.
20:07:27 [mchampion] MarkN: We could go  forward with SwA or InfoSet approach.
There may be some middle ground
20:08:13 [mchampion] We need to agree on what "attachments" means before we can
have this discussion.
20:09:37 [mchampion] MarkJ: Both options are "concrete attachments schemes" Differences
have more to do with processing model
20:11:13 [mchampion] MarkN (?): People have discomfort over "binary", "attachement"
confusion 
20:12:03 [mchampion] DavidF: New model may not be described by our existing abstract
attachements feature.  May meet some of AF requirements.
20:12:25 [mchampion] MarkJ: Abstract doc is silent on processing model, security,
other important issues.
20:12:56 [mchampion] DaveO: Do current requiremmmments cover these issues?
20:13:13 [mchampion] MarkJ: yes, and these are the highest priority
20:13:45 [mchampion] DaveO: If we need a processing model requirement, SwA still
needs more work to meet requirement.  
20:14:05 [mchampion] MarkJ: We need a processing model
20:15:09 [mchampion] MarcHadley: Not there on Friday ... need concrete attachment
feature.  Infoset approach interesting, but not clear how it would work
20:16:52 [mchampion] DavidF: There is discussion of requirements, processing model,
but there's also the "we need concrete attachment scheme NOW" issue since Rennes
meeting.  Infoset approach will take longer.
20:17:43 [mchampion] MarkN: Disputes that the Infoset approach will take longer.
Both mechanisms need work to fully specify.  Let's not rush if it's not well
specified and modeled.
20:19:10 [mchampion] Jacek: People are waiting for something because SwA changes
needed for SOAP 1.2  would be minimal.  We should rubberstamp an existing approach
in a couple of months, then consider Infoset approach.
20:20:23 [mchampion] MarcHadley: Infoset approach seems to be built on top of SwA,
let's not have two specs because we didn't do due diligence
20:21:23 [mchampion] We seem to have decided that MIME is the answer, we should
signal this to the world.  This would answer MIME vs DIME.
20:23:11 [Gudge] ACTION: Chair to investigate whether SIM approach is in scope
for XMLP WG
20:23:23 [mchampion] John: this is apples and oranges ... SwA is a concrete binding
for an attachment mechanism for binary stuff to XML envelope.  Infoset stuff is
about representing message in abstract sense.  SwA is potential binding for Infoset
model
20:25:48 [mchampion] John: Infoset model would validate our model and let us determine
whether SwA is a valid implementation of the abstraction
20:27:30 [mchampion] Anish: Agrees with John. These are two different things.
SOAP community needs concrete attachment feature
20:27:51 [mchampion] DIME vs MIME - Oracle's preference is for MIME
20:29:23 [mchampion] MarkJ: Some Requirements could be implemented by either
scheme.  Other security and processing model requirements are more than packaging
issues and would distinguish the approaches.
20:31:48 [mchampion] DavidF: In next week, let's continue convergence discussions
on list.  
20:32:12 [mchampion] Please expedite description documents if possible.
20:32:49 [mchampion] Next week we'll go through attachment feature requirements
and try to agree on broadest requirements that would encompass either approach.

20:34:10 [Zakim] WS_XMLP()2:00PM has ended

20:36:02 [davidF] rrsagent, bye
20:36:02 [RRSAgent] I see 6 open action items:
20:36:02 [RRSAgent] ACTION: Editors and W3C staff should request that we can keep
the PR namespace if nothing changes, and to get an explanation if this is not possible [1]
20:36:02 [RRSAgent]   recorded in http://www.w3.org/2003/03/12-xmlprotocol-irc#T19-30-24
20:36:02 [RRSAgent] ACTION: Gudge to chase Apache, BEA and PaulK re: nodeType [2]
20:36:02 [RRSAgent]   recorded in http://www.w3.org/2003/03/12-xmlprotocol-irc#T19-38-01
20:36:02 [RRSAgent] ACTION: Gudge should ping BEA, APache, and Kulchenko about
nodeType implementation [3]
20:36:02 [RRSAgent]   recorded in http://www.w3.org/2003/03/12-xmlprotocol-irc#T19-39-18
20:36:02 [RRSAgent] ACTION: Issue list maintainer should close 417 with notation
that issue withdrawn  [4]
20:36:02 [RRSAgent]   recorded in http://www.w3.org/2003/03/12-xmlprotocol-irc#T19-43-47
20:36:02 [RRSAgent] ACTION: Mark Jones will prepare an index for the AF requirements
doc, in order of least to most dependent on the choice of the proposals under consideraiton [5]
20:36:02 [RRSAgent]   recorded in http://www.w3.org/2003/03/12-xmlprotocol-irc#T20-03-46
20:36:02 [RRSAgent] ACTION: Chair to investigate whether SIM approach is in scope
for XMLP WG [6]
20:36:02 [RRSAgent]   recorded in http://www.w3.org/2003/03/12-xmlprotocol-irc#T20-23-11