- From: Jonathan Marsh <jmarsh@microsoft.com>
- Date: Thu, 30 May 2002 11:20:32 -0700
- To: <www-ws-desc@w3.org>
- Message-ID: <330564469BFEC046B84E591EB3D4D59C0634BE15@red-msg-08.redmond.corp.microsoft.com>
Minutes of the 30 May 2002 WS Desc telcon
Present:
David Booth W3C
Allen Brookes Rogue Wave Software
Roberto Chinnici Sun Microsystems
Glen Daniels Macromedia
Tom Jordahl Macromedia
Jacek Kopecky Systinet
Sandeep Kumar Cisco Systems
Philippe Le Hégaret W3C
Steve Lind AT&T
Kevin Canyang Liu SAP
Michael Mahan Nokia
Jonathan Marsh Microsoft
Jeff Mischkinsky Oracle
Dale Moberg Cyclone Commerce
Johan Pauhlsson L'Echangeur
Arthur Ryman IBM
Adi Sakala IONA Technologies
Jeffrey Schlimmer Microsoft
Igor Sedukhin Computer Associates
Jerry Thrasher Lexmark
Sanjiva Weerawarana IBM
Don Wright Lexmark
Joyce Yang Oracle
Prasad Yendluri webMethods, Inc.
Regrets:
Mike Ballantyne Electronic Data Systems
Keith Ballinger Microsoft
Michael Champion Software AG
Laurent De Teneuille L'Echangeur
Tim Finin University of Maryland
Mario Jeckle DaimlerChrysler Research and Technology
Dan Kulp IONA
Pallavi Malu Intel
Michael Mealling Verisign
Jean-Jacques Moreau Canon
Stefano Pogliani Sun
Jochen Ruetschlin DaimlerChrysler Research and Technology
Waqar Sadiq Electronic Data Systems
Krishna Sankar Cisco Systems
Daniel Schutzer Citigroup
Dave Solo Citigroup
William Vambenepe Hewlett-Packard
Absent:
Mike Davoren W. W. Grainger
Youenn Fablet Canon
Dietmar Gaertner Software AG
Mike McHugh W. W. Grainger
Don Mullen Tibco
William Stumbo Xerox
Sandra Swearingen U.S. Department of Defense, U.S. Air Force
Agenda:
http://lists.w3.org/Archives/Public/www-ws-desc/2002May/0253.html
1. Assign scribe.
Sandeep Kumar (David Booth for last part)
2. Approval of minutes [6].
[6] http://lists.w3.org/Archives/Public/www-ws-desc/2002May/att-0223/01-minutes-irc-02-05-23.htm
JM: 1 problem with last telcon minutes. Mike Ballantyne was marked absent but he was present. We will change it and approve the minutes.
3. Review of Action items.
PENDING 2002.05.23: GlenD to post email adding an issue: Are MEP's hard
coded in the spec, or can they be extended through
extensibility?
PENDING 2002.05.23: JeffM to write up rationale for dropping operation
overloading.
DONE 2002.05.23: Philippe to update status section of the usage
scenarios doc
PENDING 2002.05.23: Philippe, Jonathan, Waqar to work on getting the usage
scenarios published.
DONE 2002.05.23: Prasad, JeffS close the current issue, and have Prasad
open a new issue re-titled "Negative Acknowledgement"
DONE 2002.05.23: Roberto to take another round at updating his
extensibility proposal
4. Abstract Model TF. New member: Sandeep.
Discussions on AMTF findings to be included in F2F.
5. Usage Scenarios.
Publication status.
JM: Jonathan has added references to the Usage Scenario documents and send it for Director's approval for publication. Hopefully, within a few days we will publish the document.
Philippe: Yes.
6. FTF stuff
JM: Is there any F2F dial-in commitments? (none)
ACTION: Jonathan will send email to the group regarding this.
JM: Request for Martin Gudgin as the invited expert from Microsoft. In a couple of weeks, he will be replacing Keith Ballinger as the Microsoft member. He would like to observe until then. Any objection?
JK: Accept.
Jeff M.: What are the rights of the observer?
JM: He has not been very strict about it so far on Straw Polls, comments, etc. For formal votes, only members for sure.
Jeff M.: Concerns around Straw Polls.
JM: He will apply a local poicy where Martin and Keith to be counted as 1 for Straw Poll purposes. Any other concerns?
JM: None, so we will proceed as above.
JM: F2F registration ends tomorrow and he has a fairly stable list. Any logistical questions?
DB: F2F registration ends today, actually.
Tom: Is there a Joint Dinner with the Arch Group on Wed.?
Y: There is a Joint Lunch with Arch group and he is working Joint Dinner for 11th and 12th.
7. Issues. Issue list at [10]. Latest draft of the spec at [11].
[10] http://dev.w3.org/cvsweb/~checkout~/2002/ws/desc/issues/wsd-issues.html
[11] http://lists.w3.org/Archives/Public/www-ws-desc/2002May/att-0144/02-part1.html
No new issues
7a: Issue: Overloading operations
Awaiting action from JeffM?
JM: We are in agreement that Overloaded Operations will not be supported.
Jeff M. to write the rationale and we can close on this issue.
7b. Issue: Extensibility/Open Content Model
Roberto's revised extensibility proposal (sorry, message doesn't appear
in the archive). Straw poll on whether to pursue simplifications or
accept the proposal as is.
JM: Doesn't appear to be any consensus yet on what Roberto & Igor are proposing and what Sanjeeva is proposing on simplicity.
JM: At this point I think it might be a good idea to do a Straw Poll or do we need more info?
Igor: 2 proposals conflict around simplicity - either everyting is required or everything is optional. Roberto's orig. proposal was to allow both.
Roberto: The orig. proposal was not biased.
JM: Jeff's table has only 2 values: MUST or MAY. Rest is all details.
Igor: Too many combinations with additional parameters.
DB: +1 on Sanjeeva's Simplicity proposal.
Igor: Nobody wants to say No to simplification.
Igor: A global declaration is yet another parameter.
JM: Shall we do a Straw Poll?
Igor: Straw Poll on what?
DB: Straw-Poll on default True/False ..
Igor: Open content model (?)
Phillipe:
Roberto:
Igor: WSDL has to be Human readable
Jeff M.: Questions whether WSDL should be Human readable. Mostly for tools.
Igor: At least it has to Human verifiable.
JM: Let us focus on Functionality.
Roberto: What happens w/ attributes in Simpler Case?
It applies to all elememts and all attributes?
Jeff S.: Pros would be needed. Whether we need Must attributes? Not possible in WSDL 1.1.
Global extension declarations. They are all required or not. Table should reflect that.
JM: Let us breakdown functionality. Notate in file whether extension is required.
Igor: Simplification can be built with Global flag.
GD: Why can't we take the same approach as SOAP does?
JM: Straw-Poll on following three options:
a. Granularity on NameSpace level.
b. Granularity on Elements level.
c. Both the above.
Igor: Sanjeeva's proposal is on attributes, right?
DB: Simplicity has multiple meanings: (a) Writing is simpler (b) Semantically it is simpler
JS: We should go for Semantically Simpler
JS: What would be the processing model for Imports?
GD: Process all imports first.
JS: Makes sense.
GD: Namespace is potentially a mistake.
Igor:
GD: Closer to SOAP, better
Igor:
GD: QNames represent extensions and they are required
Roberto:
GD: Describes the proposal using an example of security extensions with 5 elements QNames are MustUnderstood and all the 1000 elements are MustUnderstood as well.
[Scribe changes to David Booth]
GlenD: You can make up a token, and make it required, that implies that you must also understand several other elements.
DBooth: This is an application level contract.
Igors: I don't think Roberto's proposal is opposing what we're discussing. I think we're discussing how to indicate requiredness
... Do we need granularity per element?
... How do we declare? By namespace? By qname? etc.
JM: ANd also what do we do about attributes?
... There isn't an easy way to annotate individual attributes.
GlenD: Yes there is.
ACTION: GlenD to write up his ideas re: extensibility
DBooth: The more I think about Glen's suggestion, the more I like it. I don't think it would be good to use a namespace to indicate blanket "requiredness".
JM: We could choose either the WSDL required element or attribute.
GlenD: Or both.
JM: I'd like to see examples of likely extension scenarios. E.g., an extension element that contains within it, different namespaces.
7c. Issue: Optional parts in <message/>
Thread starts at [22]. What is the status of this issue?
[22] http://lists.w3.org/Archives/Public/www-ws-desc/2002May/0004.html
JeffS: We decided earlier that we don't want to further extend WSDL-specific parts for functionality.
JM: Are the other issues in the draft or on our issues list? Are we actively discussing them?
JeffS: We decided we will not add optionality to message parts.
... Also we will leave the message construct alone.
Prasad: I had some things to say on these issues. Most of the WS would come from existing implementations in organizations. There are lots of existing schemas that people want to reuse. There are constructs now that use different schema features. The good thing about the message construct is that it allows multiple parts to be assembled into one entity.
... You want to be able to have attachment that are optional in messages.
... At least you should have a place holder.
JM: Do we have a process for deferring an issue?
... We decided we will not do optional parts, so we just need to remember that.
JM: So I guess we're done on that.
JeffS: I think we should mark these two off the list.
Prasad: We agreed we will not remove WSDL:Message from 1.2.
... To really address this issue we need to decide whether "message" should go.
JM: Can we decide this without deciding 2.0 issues first?
Prasad: We could defer this pending resolution of the other issue. (Beyond 1.2)
GlenD: General note: I'm a little concerned by the group's tendency to steam-roller past several of these issues without considering them. I'd like this language to, as quickly as possible, be something that we want to carry into the future.
... So I don't think we should too blithely sweep issues off to 2.0.
... I.e., If we later at a F2F come up with something good, I don't think we should be prevented from discussing it.
JM: As chair I need to be careful to avoid covering ground that was already covered, but if there's new information, that's different.
... So to the best of our knowledge now, we don't see the value of changing the "message" construct, so we'll leave it as is unless new info comes up.
Prasad: The spec describes things as optional, but isn't clear. We need to fix that.
... It is inconsistent.
ACTION: Prasad to raise an issue of spec inconsistency about optional parts.
7d: Issue: Negative Acknowledgement
Issue recast at [12].
[12] http://lists.w3.org/Archives/Public/www-ws-desc/2002May/0232.html
TJordahl: If a notivfication msg is only one way, then there can't be a response, so I don't understand why this is an issue.
JM: Yes, that's why Prasad retitled the issue.
TJordahl: If you want to know if an operation failed, then you should get a response that may be empty if there's no problem.
JeffS: Negative Ack is a pretty common protocol paradigm.
Philippe: The definition of "fault" is not clear enough in WSDL. Is it an app fault? A protocol fault? If I don't understand an extension, is it a fault?
JeffS: If Soap headers are not understood, then it is a fault.
... WSDL was not clear. We should be clearer.
DBooth: I agree with Philippe. ALso, the purpose of NAck is to save bandwidth. That purpose is omportant for wire protocols, but may be less relevant at the Web Service level.
JM: We need a concrete proposal to discuss.
TJordahl: I think adding another MEP (NAck) would make WSDL more complicated, and I wouldn't favor it. It may be one of those things that nobody implements.
... It doesn't make WSDL "more friendly", it just makes it "more".
JM: I have seen Prasad's interest. ARe there others interested in it also?
(No response)
JM: Straw poll: SHould we close the issue or entertain a proposal?
JM: JeffS and Prasad in favor continuing the issue; all others were for closing the issue.
JM: Objections to closing?
Prasad: I think people will see later that they'll need it, but I'm willing to close it.
JeffS: I'm willing to go along with closing it also.
JM: It seems like an interesting feature, but most people view it as not in the minimal set that we need.
[Meeting adjourned]
[Chair reminds members to start discussion of remaining issues for next week:
7e. Issue: remove solicit-response and output-only operations?
7f. Issue(s): Non-SOAP HTTP Binding
7g. Issue: Should Operations permit alternate and multiple responses?
7h. Issue: portType extensibility.
7i. Issue: service type.]
Summary of Action Items:
2002.05.23: GlenD to post email adding an issue: Are MEP's hard
coded in the spec, or can they be extended through
extensibility?
2002.05.23: JeffM to write up rationale for dropping operation
overloading.
2002.05.23: Philippe, Jonathan, Waqar to work on getting the usage
scenarios published.
2002.05.30: Jonathan will send email to the group regarding this.
2002.05.30: Prasad to raise an issue of spec inconsistency about
optional parts.
Received on Thursday, 30 May 2002 14:21:45 UTC