W3C

Web Service Description WG TelCon
12 Dec 2002

See also: IRC log

Attendees

Present:
Erik Ackerman Lexmark
Mike Ballantyne Electronic Data Systems
David Booth W3C
Allen Brookes Rogue Wave Software
Roberto Chinnici Sun Microsystems
Glen Daniels Macromedia
Youenn Fablet Canon
Dietmar Gaertner Software AG
Tom Jordahl Macromedia
Jacek Kopecky Systinet
Sandeep Kumar Cisco Systems
Philippe Le Hégaret W3C
Amelia Lewis TIBCO
Steve Lind AT&T
Kevin Canyang Liu SAP
Lily Liu webMethods
Jonathan Marsh Chair (Microsoft)
Jeff Mischkinsky Oracle
Dale Moberg Cyclone Commerce
Don Mullen Tibco
Arthur Ryman IBM
Adi Sakala IONA Technologies
Jeffrey Schlimmer Microsoft
Jerry Thrasher Lexmark
William Vambenepe Hewlett-Packard
Sanjiva Weerawarana IBM
Prasad Yendluri webMethods, Inc.
(irc) Barbara Zengler DaimlerChrysler Research and Technology
Regrets:
Steve Graham Global Grid Forum
Jean-Jacques Moreau Canon
Waqar Sadiq Electronic Data Systems
Igor Sedukhin Computer Associates
William Stumbo Xerox
Joyce Yang Oracle

Chair: Jonathan Marsh

Scribe: Don Mullen

Contents


1. Assign scribe : Don Mullen

2. Action Items

Scribe: Paco - pending
... Roberto - pending
... Marsh - done
... Glen/Paco - pending
... Don - pending
... Marsh - pending draft
... Arthur - pending

<Arthur> http outbound ops - under internal review

Scribe: Glen - pending
... Marsh - done
... JeffS - pending
... Kevin - done
... Dbooth - done

4. Administrivia

<dbooth> Please register for the January F2F: http://cgi.w3.org/Register/selectUser.pl?_w3c_meetingName=WSAWG_WSDWG_200301

Scribe: Register early - prices increasing.

<dbooth> Hugo, Philippe and I opted to save our organizations money, so we are staying at the Resort for $139/night: http://www.resortsuites.com/ No high-speed internet there, but it does have a golf course.

<jeffsch> +1 to end face to face meeting at noon on Wednesday

Scribe: F2F meeting will end at noon Wed

<jeffsch> scribe, where is the may face to face?

Scribe: Week of May 12 in France - possible July in Toronto.

<GlenD> jeffsch: Rennes, France

Scribe: March 3-4 F2F in Cambridge, MA - Tech Plenary Session that week
... F2F Dec 26th canceled / Jan 2 also canceled (possible TF use)

5. Publication issues

Marsh: will not make Dec date for publication. Hopefully draft by early Jan that we can review and initiate publication process.
... W/web S/service issue : ask architecture group for glossary. WS-Arch uses 'Web service'.
... should we go ahead now and put in our draft as such?

Roberto: Vote count in the architecture group?

Marsh: Report was evidently strong in favor for 'Web service'.

Glen: Vote 13-6

<Arthur> Web service is fine with me (I even voted for it!)

Marsh: propose to use architecture group's advice -- 'Web service' - group agreed.

<Arthur> yeah us!

6. New Issues - none

7. Output operations

Scribe: Youenn reviews his view: http://lists.w3.org/Archives/Public/www-ws-desc/2002Dec/0039.html

Youenn: argues for factoring out inbound and outbound -- separate from assigning a MEP

JeffS: Could support to go out for primitive input and output only - push forward with MEP work.

Youenn: Clarification - have input and input-output still. Add 'inbound' and 'outbound' - factored out.

Glenn: Would you keep the input and output element declarations?

Youenn: Problem with input and output elements -- element order is important now.
... Order of elements should not be taken into account.

Glen: Agree, but certain elegance to leaving as is and have MEP in addition.

Sanjiva: More complex examples make order more imprecise.

Glen: Problem with sliding into choreography....

Sanjiva: MEP defines messages in/out -- flow is choreography.

Glen: Thought problem with output -- how to make the connection in interoperable way.

Sanjiva: Subscription/notification example.

Marsh: Envision any MEPs in future where identifying the role of web service is difficult?

Glen: Composition - i.e. MEP at transport level mapped to application level - layering things makes possible node
... involved with more roles

Marsh: Raise issue of TCP binding - Sanjiva not in favor -- comments?

Sanjiva: Would like to find a way to sort out output in detail -- not 'punt' with just publishing note.
... proposal was keep output operations as is - and just publish note. Need more than current status quo (not defined).
... Microsoft kind of like 'call-back' -- TIBCO is more publish to information bus.

<jeffsch> Microsoft has not committed to a specific meaning of output operations beyond saying that they are messages sent by the service.
... We view this as comparable to the spec's commit to the 'meaning' of input operations.

<GlenD> I agree with that interpretation, Jeff, though I think it's not adequate. :)

JeffS: Why are there a problem with 'output' and not 'input' -- 'input' is just as underspecified.

<Arthur> the WSDL 1.1 spec does not define any bindings for outbound ops

<alewis> Agreed: this is a binding issue.

<dbooth> agrees that the problem is clearly a binding issue.

<jeffsch> Of course, willing to clarify output _and_ input direction as needed.

<DonM> completely agrees -- let's move forward with MEPs -- then more details on binding.

<sanjiva> what clarification do you seek for inbound operations?

Kevin: Abstract at design-time more important -- deployment time binding becomes important.

Marsh: Question for Kevin: problem with defining roles at an abstract level?

Kevin: Must have capability at design-time to describe message -- how bound at deployment separate. Unclear how MEP would help.

Marsh: Seems like direction last week (TCP note + leave status quo) not going to work.

TomJ: What is the good resolution here (like asked last week)? WSDL should fully describe how to use output or input operation.
... if people using WSDL 1.1 and happy with that -- can we leave as is?

Marsh: So what clarifications do we need to make more interoperable?

Don: Clarification - those using WSDL 1.1 output operations making due -- definitely need more. MEPs a great step forward.

Glen: Need more

Scribe: Don/Amy: Agree, but MEPs go a long way to helping aleviate the problem.

Arthur: WSDL 1.1 spec - defines these things -- only defines bindings for one-way and request/response. Allows for more....
... Missing piece is in the binding.
... IBM trying to get concensus on this - bring forward. Not fundamentally wrong.

Jacek: Thought TCP note would provide solicit-response binding...

<youenn> +1
... .. to arthur

Arthur: HTTP binding could do notification -- could specify more completely.

Glen: Gets into application level

<sanjiva> Jack: IMHO the spec shouldn't include stuff for which it doesn't define normative bindings for. That's what extensibility is for .. it should not preclude the desired functionality, but it should not leave it in.

DBooth: What about task force to address this issue in detail?

JeffS: MEPs interesting -- originally just clarifying WSDL 1.1 spec. MEPs perhaps extension.

Sanjiva: We are now taking on more.

JeffS: Should we close output issue and more forward with MEPs in more detail?

<Arthur> see http://www.w3.org/TR/wsdl#_porttypes "It is expected that specifications that define the protocols for Solicit-response or Notification would also include WSDL binding extensions that allow use of these primitives.

Amy: MEP issue encompasses both.

JeffS: Makes life a little more complicated, but could live with that.

Marsh: Move forward more on MEPs, but also continue with Arthur's work as well.
... MEPs -- 1) could add syntax to WSDL and add information via uri
... 2) specify output and output-input more clearly -- define specific MEPs to include in spec.

Glen: argues that separate MEPs complicates what is now simple.

Marsh: Should we write MEP specifications for output and output-input?

Scribe: JeffS/Amy: Should include all -- Amy preference to start with what we know more about input/output.

Marsh: Would help to get some of the details of MEPs specification under way.

Amy: Don's action about making SOAP general helps get input-output.

Glen: Isn't output and output-input reverse -- and little work.

Amy: No. Resist idea that these are simply reverse.

<jeffsch> I think Tibco\Amy doesn't accept that output-input should always be the reverse of request-response. Address this by defining > 1 MEP for an operation with output followed by input?
... ... Or > 1 binding to that type of operation?
... One MEP would be ==1 output followed by >= 0 inputs...
... ... the other MEP would be ==1 output followed by ==1 input?

<GlenD> Task Force
... MEP SWAT team

Scribe: General discussion of importance of defining multiple MEPs depending on the programming model.

Sanjiva: Concern about two MEPs that aren't fully specified -- no binding.

JeffS: Proceed with Arthur or TCP specific examples of how to do this.

<dbooth> I again propose a task force.

<GlenD> +1! Task Force!

<bzengler> i think a task force would be a good idea.

GlenD: Telcon with smaller group also would be beneficial.

JeffS: Like pushing foward MEP idea as top priority -- more clarity and examples. Have some questions.

<jeffsch> Willing to make output direction ops a dependent issue on the MEP design issues and move the latter to the front of the agenda.

Marsh: Hoping email / regular telcon would address this.
... Will create a telcon for Tuesday at 8am next week to discuss specifics of pushing forward on MEPs.

<bzengler> a telcon for a smaller group or for the whole WG?

<GlenD> Interested Conspirators :)

<bzengler> ok, thanx :)

Marsh: Other items that would help? Arthur's work. Others? Note Arthur on vacation.

Summary of Action Items

  • 2002-11-12: Glen and Paco to chase the Global Grid Forum WRT services implementing a single portType
  • 2002-11-12: Paco will write two options for naming faults: schema vs WSDL.
  • 2002-11-12: Roberto will try and come up with another proposal for eliminating message, the discussion goes to email or the next f2f.
  • 2002-11-21. Don Mullen to detail changes/addition necessary to unify SOAP and WSDL MEPs.
  • 2002-11-21: Jonathan to refer R120 text to TAG, referencing TAG issue fragmentinXML-28, when that text appears in the draft.
  • 2002-12-05: Arthur to propose an HTTP binding approach for output only
  • 2002-12-05: Glen to write up a description of the issues surrounding property description in WSDL.
  • 2002-12-05: JeffS to see what would be needed to publish his TCP binding on output only as a Note.
  • 2002-12-12: Jonathan to set up a telcon for 11AM EST Tuesday - Topic: MEPs.

  • David Booth
    dbooth@w3.org
    $Date: 2002/02/19 16:35:31 $