- From: Sakala, Adinarayana <Adi.Sakala@iona.com>
- Date: Mon, 21 Jul 2003 12:22:15 -0400
- To: <www-ws-desc@w3.org>, <w3c-ws-desc@w3.org>
- Cc: <jmarsh@microsoft.com>
W3C Web Services Description Teleconference 7/17/2003
Minutes of Meeting
--------------------------------------------------------------------
Present:
Erik Ackerman Lexmark
Mike Ballantyne Electronic Data Systems
David Booth W3C
Allen Brookes Rogue Wave Software
Roberto Chinnici Sun Microsystems
Glen Daniels Macromedia
Alan Davies SeeBeyond
Dietmar Gaertner Software AG
Steve Graham Global Grid Forum
Tom Jordahl Macromedia
Jacek Kopecky Systinet
Philippe Le Hégaret W3C
Amelia Lewis TIBCO
Steve Lind AT&T
Kevin Canyang Liu SAP
Lily Liu webMethods
Jonathan Marsh Chair (Microsoft)
Dale Moberg Cyclone Commerce
Bijan Parsia University of Maryland MIND Lab
Arthur Ryman IBM
Adi Sakala IONA Technologies
Jeffrey Schlimmer Microsoft
Bryan Thompson Hicks & Associates (DARPA)
Jerry Thrasher Lexmark
William Vambenepe Hewlett-Packard
Sanjiva Weerawarana IBM
Umit Yalcinalp Oracle
Prasad Yendluri webMethods, Inc.
Regrets:
Sandeep Kumar Cisco Systems
Ingo Melzer DaimlerChrysler
Jean-Jacques Moreau Canon
--------------------------------------------------------------------
1. Minute Taker: Adi Sakala
--------------------------------------------------------------------
2. Approval of minutes:
- July 10 telcon [.1]. Alan Davies sent belated regrets. Jonathan
asked for two AIs to be recorded [.2].
[.1] http://lists.w3.org/Archives/Public/www-ws-desc/2003Jul/0052.html
[.1] http://lists.w3.org/Archives/Public/www-ws-desc/2003Jul/0054.html
[APPROVED]
--------------------------------------------------------------------
3. Review of Action items [.1].
2003-03-13: Don will write a proposal for annotating schema
with part information.
[PENDING]
2003-03-27: Philippe write up a proposal for embedding binary
data types in schema
[PENDING] - will be ready for F2F
2003-05-13: DaveO to send a motivating example for R131.
[PENDING]
2003-05-13: Jeffsch, Sanjiva, Glen, Umit, JJM to come up with
a proposal to get rid with the message construct,
and add programming hints.
Jeff: sanjiva already has a proposal and waiting for consenses
among the smaller group of people.
JM: would like to see progress by F2F
[PENDING]
2003-06-12: Jacek to synthesize the different approaches
to solving issue 64.
[DONE]
DONE [.2] 2003-07-03: Arthur to figure out which validation mode our
schema should specify on xs:any.
2003-07-10: DBooth to reconcile his terminology with that
in the requirements doc.
[DONE]
David: it is quite consistent with the core spec, but consistent with the schema.
DONE [.3] 2003-07-10: Jonathan to send a summary of the @serviceGroup
idea to the list.
[.1] http://www.w3.org/2002/ws/desc/#actions
[.2] http://lists.w3.org/Archives/Public/www-ws-desc/2003Jul/0068.html
[.3] http://lists.w3.org/Archives/Public/www-ws-desc/2003Jul/0055.html
--------------------------------------------------------------------
4. Administrivia
a. July FTF logistics [.1], registration [.2]
b. Sept FTF logistics [.3] (need registration page)
[.1] http://www.w3.org/2002/ws/desc/3/05/f2fJulyLogistics.htm
[.2]
http://cgi.w3.org/Register/selectUser.pl?_w3c_meetingName=WSAWG_WSDWG_20
0307
[.3] http://www.w3.org/2002/ws/arch/3/07/f2fSeptLogistics.html
Jerry: will there be any joint meeting on wednesday?
JM: it is a open day, nothing on agenda for now.
JM: will work on agenda and will send out a draft in next couple of days.
ACTION: dbooth to create/link logistics page for Sept F2F
------------------------------------------------------------------
5. Task Force Status.
a. Properties and Features (dormant)
b. Patterns
c. Attributes
d. QA & Testing
------------------------------------------------------------------
6. New Issues. Merged issues list [.1].
- NONE
[.1]
http://dev.w3.org/cvsweb/~checkout~/2002/ws/desc/issues/wsd-issues.html
------------------------------------------------------------------
7. Open content model - validate "strict" [.1].
[.1] http://lists.w3.org/Archives/Public/www-ws-desc/2003Jul/0068.html
Arthur: gudge pointed out that official XML Schema uses lax.
Prefers to validate as much as possible and schemas
should be provided.
RESOLUTION: Agreed to use LAX as validation.
Editors ToDo: Editiors to update specs and schema with this consenses.
--------------------------------------------------------------------
8. Polishing off single service per interface
Terminology [.1] Diagram [.2, .3]
Alternatives: @targetResource
@serviceGroup [.5]
- Naming the service resource [.4]
[.1] http://lists.w3.org/Archives/Public/www-ws-desc/2003Jul/0018.html
[.2] http://lists.w3.org/Archives/Public/www-ws-desc/2003Jun/0064.html
[.3] http://lists.w3.org/Archives/Public/www-ws-desc/2003Jul/0019.html
[.4] http://lists.w3.org/Archives/Public/www-ws-desc/2003Jul/0008.html
[.5] http://lists.w3.org/Archives/Public/www-ws-desc/2003Jul/0055.html
Glen: serviceGroup semantics seem identical to targetResource semantics
David: +1 to Glen's comment: serviceGroup semantics seem identical to targetResource semantics
Bijan: the idea was to add further semantics if we use serviceGroup
Glen: prefer removing targetResource and serviceGroup and make wsdl more simple.
It is so fuzzy and we dont know what we are doing with this.
Umit: targetResource is more accurate than serviceGroup.
David: agrees with Umit that serviceGroup is a half-assed solution to indicating relationships between services
- bijan notes that serviceGroup + and assertion that the memberfs of some group manipulates the same resource is *exactly* the same semantics as targetResource. How is it not?
David: didn't hear: identify *what* with the same URI?
Jacek: agrees with bijan, and everyone will be able to use their assertions, ontologies, and not redefine what "manipulation" means
Tom: +1 to abandoning serviceGroup or targetResource
JacekK: +1 as well
sanjiva: are we still keeping service/@interface? (i.e., single interface/service)
JacekK: that's not being discussed, AFAIK
JM: Glen is asking if either targetResource or serviceGroup is required,
David: We abandon the idea of both as it can be achieved out side wsdl by other specs and
inside wsdl by extensions.
Glen: we are not ready or concrete to propose in the area of state management and sort of
relationship between services.
jeffsch: +1 to Macromedia\Glen that we may be premature in trying to standardize one of these
bijan: +1 to that too
JacekK: says that, it's in one of the past minutes that we might drop targetResource
and wait until someone proposes something concrete and not vague
Bijan: I rather debate to have concrete proposal than vague description.
JM: are we talking about dropping diagram.
Glen: diagram is still useful as it defines some sort of relationship.
Umit: The whole proposal is not about state management but about defining
something with a URI so that we know we are accessing the same thing (resource).
This helps identifying the identities that represents the same resource and will
be useful further to define a concrete state management.
Bijan : We need more clear text if we want to keep it.
I dont understand how to use targetResource and what are the problems associated with it.
Glen: if targetResource is just a URI, can we achieive it by just comparing URI's
David: The meaning of it is completly dependent on the application.
The situation is similar if we provide or not provide targetResource.
It doenst use it in interoperating.
Umit: if we get rid of targetResource, Is the single interface per service still Acheived?
Should we remove targetResource attribute? (Straw Poll)
Yes - 24
No - 3
Obstain -
JM: Would anybody object to recording removing TR attribute?
Group has consenses.
RESOLUTION: Remove TargetResource attribute
Discussion about Diagram:
David: New diagram can be found at, http://lists.w3.org/Archives/Public/www-ws-desc/2003Jul/0100.html
David: Service will have only one interface, but a single interface can be associated with
multiple services.
JM: do we really need this diagram at all any more?
JM: main purpose of diagram without targetResource is just showing what you see reading XML.
Sanjiva: diagram is useful
JM: we are looking at a graph to explain the relationship.
Options:
1. Leave the existing diagram and modify to remove the resource from the diagram
2. remove the diagram completly.
3. replace the diagram with the latest diagram that David submitted and it will
also have resource removed.
Straw Poll:
Option 1:1
Option 2:24
Option 3:
Obstain:
Overwhelming support for Option 2.
RESOLUTION: Get rid of Diagram.
bryan: I have a hard time seeing how WSDL can drop the reference to "resource" and address both web and web services architecture.
sanjiva: WSDL 1.1 did not refer to "resource" in any strong sense IIRC
JacekK: bryan, this is IMHO related to issue 64 and my last email on this topic
umit: +1 to bryan. If WSDL 1.1 is enough, then what are we doing here :-)
sanjiva: Hey I suggested that earlier ;-)
sgg: lets go with WSDL 1.1, add open content and portType inheritance and I would be very happy
Terminology:
David: we dont think we have a problem with terminology as we got rid of targetResource attribute.
Suggestion to Editiors: add more explainatory material as part of terminology.
Naming the service Resource Proposal:
ACTION: David Booth to reply to Anne Manes asking what would be the diff between her proposed service URI and the service QName: http://lists.w3.org/Archives/Public/www-ws-desc/2003Jul/0008.html
--------------------------------------------------------------------
9. Binding enhancements
Sanjiva proposes [.1] to:
1) Drop @interface from binding, since now in service.
2) Allow inlining interface-wide binding within a port and
making binding optional.
3) Define default binding (SOAP doc/lit).
4) Dealing with operation specific SOAPActions.
Kevin proposes to [.2]:
1) Allow reuse of parts of bindings through a BindingDetails
element and corresponding references.
Consolidated proposal [.3]
Amy's feedback [.4]
[.1] http://lists.w3.org/Archives/Public/www-ws-desc/2003May/0046.html
[.2] http://lists.w3.org/Archives/Public/www-ws-desc/2003Jan/0068.html
[.3] http://lists.w3.org/Archives/Public/www-ws-desc/2003Jul/0017.html
[.4] http://lists.w3.org/Archives/Public/www-ws-desc/2003Jul/0051.html
sanjiva: Does everyone remember the proposed binding simplifications?
Amy: gave feedback on sanjiva's proposal.
Tom: without a service it makes difficult to know more information about binding.
Tom: I think I agree with Amy - Having the interface on the binding is very handy
Glen: Binding only makes sense with a Port.
sanjiva: you can codegen with an interface only if you want (and u can generate only the interface). If you wnat to codegen a full stub then you *must* have a <service>
Tomj: Is that true in the Jax-RPC sense?
sanjiva: TomJ: Yes, there's no codegen u can do with interface+binding only ..
Tomj: OK, but I am still concerned that if I have a binding in hand, its just going to be very unpleasant to discover what interface this binding it talking about.
David: thinks he's hearing that if a binding mentions an operation, then it would have to look up what interface that operation belongs to. Can an operation belong to more than one binding?
sanjiva: david, you can define any operatoin any # of times; did I misunderstand?
bryan: can an operation belong to more than one binding? - yes by the narrative in this section.
sanjiva: operations don't "belong" to bindings
Roberto: Sanjiva, In JAX-RPC to generate a stub you only need a binding.
JacekK: in WSIF you only need an interface
sanjiva: yes but in the sense that the address is not needed right?
Roberto: yes, the address doesn't need to be there
sanjiva: Roberto, does jax-rpc spec how a wsdl 1.1 file w/o a <service> should be codegen'ed or does it say 'start @ service and do this'
Roberto: You start with services. Stubs by themselves are unusable, they have to be obtained via a service. But the mapping from a binding to a stub is fully specified without referring to services/ports.
Roberto: So in theory you can create stubs from a WSDL 1.1 without services, but you can't use them (portably).
JacekK: you can start with a binding, and the promise that somebody will give you the endpoint information later
JacekK: that's a common case, in fact
JacekK: the "somebody will give you the endpoint information" is not very interoperable, but common
Jeff: Operations are now QNames and there is no confusion about which operation
you are binding to.
Amy: what happens in Inheritance.
Sanjiva: All operations have to be globally unique.
There is no inheritance, only union of operations.
Each portType has to have a unique targetNamespace.
Glen: should we make specifying a interface optional.
when it is not specified then it allows to reuse bindings.
Amy: what happens when you leave out an operation in the binding.
in that case we have to use defaults which causes big problems at runtime.
Glen: making it optional is perfectly valid as it is very much same to how you
want to validate the wsdl.
Sanjiva: want to know what others think on this proposal. i have heard what Amy thinks.
bryan: what is the history with respect to using UML diagrams and explanatory narrative to make concrete proposals for discussion?
JM: we have no history of UML :-)
--------------------------------------------------------------------
Summary of Action Items:
2003-03-13: Don will write a proposal for annotating schema
with part information.
2003-03-27: Philippe write up a proposal for embedding binary
data types in schema
2003-05-13: DaveO to send a motivating example for R131.
2003-05-13: Jeffsch, Sanjiva, Glen, Umit, JJM to come up with
a proposal to get rid with the message construct,
and add programming hints.
2003-07-17: David Booth to create/link logistics page for Sept F2F
2003-07-17: David Booth to reply to Anne Manes asking what would
be the diff between her proposed service URI and the service QName.
Received on Monday, 21 July 2003 12:22:23 UTC