- From: Jonathan Marsh <jmarsh@microsoft.com>
- Date: Thu, 21 Nov 2002 09:44:27 -0800
- To: <www-ws-desc@w3.org>
Adding attendance list and more detail on action items...
Present:
Allen Brookes Rogue Wave Software
Roberto Chinnici Sun Microsystems
Glen Daniels Macromedia
Martin Gudgin Microsoft
Jacek Kopecky Systinet
Sandeep Kumar Cisco Systems
Amelia Lewis TIBCO
Steve Lind AT&T
Kevin Canyang Liu SAP
Jonathan Marsh Chair (Microsoft)
Dale Moberg Cyclone Commerce
Jean-Jacques Moreau Canon
Don Mullen Tibco
Jeffrey Schlimmer Microsoft
Igor Sedukhin Computer Associates
Jerry Thrasher Lexmark
Sanjiva Weerawarana IBM
Don Wright Lexmark
Joyce Yang Oracle
Prasad Yendluri webMethods, Inc.
Barbara Zengler DaimlerChrysler Research and Technology
Regrets:
David Booth W3C
Youenn Fablet Canon
Dietmar Gaertner Software AG
Steve Graham Global Grid Forum
Philippe Le Hégaret W3C
Arthur Ryman IBM
William Vambenepe Hewlett-Packard
--------------------------------------------------------------------
3. Review of Action items.
DONE 2002-09-10: Gudge to provide summary of using W3C XML Schema
to wrap other type systems at an appropriate level
of abstraction.
DONE 2002-10-17: Gudge and A. Lewis to make a proposal for
describing xs:import and (apparently) xs:schema
as "required" extensibility elements.
OBSOLETE 2002-11-07: Jonathan to dig up use cases for various kinds
of equivalence in archives, etc.
DONE 2002-11-11: Part 1 editors to add to an appendix outlining the
areas requiring transitional documentation.
Specific text TBD.
DONE [.1] 2002-11-11: David O will post a summary of URN and fragment
identifiers with namespace issues to TAG.
PENDING 2002-11-11: Arthur will submit targetNamespace/simplifiedNun
text to editors to be included in part 1 as
non-normative appendix, including a note
regarding non backward compatibility with wsdl
1.1.
PENDING 2002-11-12: Paco will write two options for naming faults:
schema vs WSDL.
PENDING 2002-11-12: Roberto will try and come up with another proposal
for eliminating message, the discussion goes to
email or the next f2f.
PENDING 2002-11-12: Marsh to contact the XMLP WG (and later the wider
public) speaking about the issues around removing
the use attribute.
DONE 2002-11-12: Gudge to ammend WSDL schema to use substitution
groups for extensibility rather than wild cards.
[Long discussion]
Marsh: Would prefer comments in Schema rather than prose in spec
describing the Schemas.
Gudge: No worries.
[We will add prose in spec.]
PENDING 2002-11-12: Glen and Paco to chase the Global Grid Forum
WRT services implementing a single portType.
PENDING 2002-11-13: Glen to write the 2 proposals for specifying
binding
properties and send them to the list (1 = add
<*:property> to <binding>, 2 = move protocol
binding stuff including properties to ports.)
[.1] http://lists.w3.org/Archives/Public/www-ws-desc/2002Nov/0050.html
Admin
=====
New drafts for review by next Friday
New issues
==========
* Prasad: express optionality from SOAP headers (from WS-I)
* Youenn: open content model and extensibility
MediaType registration
======================
Marsh: minutes from f2f incorrect. Recall:
* Yes, register mediatype; follow TAG process
* Add issue to spec appendix: text/xml or application/wsdl+xml
jjm: this is what I recall as well. Part 1 contains new
mediatype appendix, with ednote as you have just highlighted.
7) WSDL URI
===========
Marsh: notify TAG we're doing something broken? new TAG issue?
simply tell TAG we're done?
JeffS: Arthur not on the call.
Marsh: think this is exactly issue 28 (fragment identifier)
Marsh: suggest send message to TAG when we incorporate Arthur's
text into spec. Reasonnable?
No objections.
8) Alternative schema languages
===============================
Marsh: just read text from Amy. Jack?
Jack: Amy's examples are misleading (or broken)
Amy: why should type and element be reserved for XML Schema?
Gudge: difficult, because in RelaxNG, don't have QNames
Amy: this is why I did it this way
Jack: would be simpler and easier if make them XML Schema specific
Amy: example of such schema that does not support?
Sanjiva: people have been asking for MIME explaination
Glen: link with SOAP w/ attachment?
Sanjiva: two independent things.
Amy: can't use WSDL with other schema language
Sanjiva: yes you can, with new attribute
Jack: what if import RelaxNG and XMLSchema, how would we know
which schema we point to?
Amy: described in proposed text
Gudge: proposal needs to be refined
Jack: ok, likes the proposal overall
Marsh: include in Dec 6 draft?
Snajiva: must support multiple schema languages. In addition, use
type element for that.
Amy: impacted by removal of message element
JeffS: add to the spec, be assertive we will support other schema
languages
Straw poll:
Option 1: reuse element and type attribute
Option 2: used qualified attributes for other schema languages
Option 3: reused element and type attribute for schema languages
described in XML only
Option 1: 0
Option 2: 5
Option 3: 10
Abstentions: 5
Marsh: option 3
Jack: how to solve Amy's issue?
Gudge: new issue; solve later.
Amy: not permitted; open question.
Marsh: incoporate text for Dec 6 draft, with Jack's issue?
Gudge: yes, will make necessary changes to Amy's text
9) SOAP headers independent of SOAP modules
===========================================
Marsh: just noticed JJM posted mail on XMLP; any action?
JJM: little reaction from XMLP
Glen: do you think we need this? I don't
JJM: fine with status quo
No objection to status quo.
Marsh: won't track this thread any more
10) @@@
=======
Marsh: waiting for revised proposal from Roberto
Roberto: keep it on email for a while
11) Output operation
====================
Marsh: data trader operation. How to move that forward?
Amy: TIBCO feels tied up with feature support and Don's named MEPs.
12) Revised proposal from Amy
=============================
Marsh: (for Barbara and Prasad) broke feature proposal into
manageable chunks. Had trouble at highler level of abstraction. Amy?
Amy: write of where I think we should be going. Contains a number
of clarifications. Low enough in impact and flexible enough. New
element as child of binding (protocolBinding). May contain
propertyConstraint or featureBindings, which themselves may
contain propertyConstraints. Impact low, but enough functionalities.
Glen: mechanims like this required. Like some aspects. Overlap
some other work: protocol binding that wraps up in URI both
binding to SOAP and to particular SOAP protocol binding.
PropertyConstraints at SOAP wide level. We're missing ability to
separate SOAP binding from underneath protocol binding.
JJM: also would like clearer separation between message
serialization (for example XML SOAP) and underlying protocol
binding (for example SOAP over HTTP or SOAP over Email).
Glen: support particular property constraints within particular
operations.
Marsh: continue to email; amend Amy's proposal
JJM: may provide a list of issue brought out so far, so can
compare the proposals and see missing features. No time for next
week; maybe week after that.
13) MEP support in operation
============================
Marsh: Don's proposal
JJM: no time to read it
Glen: like it!
Don: link between operations and MEPs
Jack: WSDL MEPs are different from SOAP MEPs
Sanjiva: agree, but don't think this is what Don's proposing
Glen: MEPs should move from XMLP to WS-Arch
JeffS: operations today are a sequence of messages?
Sanjiva: yes
JeffS: so, if had more than one, whether MEP, would be a sequence?
Sanjiva: one possible semantic; would not recommend that
Glenn: branching?
JeffS: if operations are not primitive, why have multiple
operations per porttype?
Sanjiva: MEP would describe message involved with that specific
interactions. Input output not enough
Jack: two MEPs in WSDL currently. If go further, will do
choreography work
JJM: wouldn't choreography build on that fundation?
Amy: yes, current two MEPs are not sufficient
Glen: need define boundary with choreography; but can still
explore the issue
Amy: not promote definition of MEPs that turn into choreography,
but should be able to express, for example, request-response MEPs
with different semantics.
Marsh: what next? how roles work?
Jack: how SOAP MEPs are WSDL MEPs.
Sanjiva: why SOAP dependency? Don is simply suggesting message
flow semantics.
Jack: pb with example; Don uses SOAP MEP URI.
Marsh: what's benefit if can't reuse SOAP URIs?
Sanjiva: e.g., TIBCO and Microsoft have different semantics for
output.
Glen: need to connect with SOAP MEPs. Could refer to abstract
MEPs. Common framework for all spec.
Sanjiva: SOAP MEPs would come at the binding level.
Marsh: rewrite SOAP req-resp MEP in a WSDL friendly way. Better
if also SOAP friendly. Can't rely on XMLP or WS-Arch. Should come
up with example. Any one with time and interest?
Sanjiva: current proposal simply extends current model.
Amy: Marsh would like abstract req-resp pattern. Would offer to
do this, but not sure right person.
Marsh: can't unify SOAP and WSDL MEPs. Need evidence.
Don: will detail changes/addition necessary to unify SOAP and
WSDL MEPs.
Adjourn.
Received on Thursday, 21 November 2002 12:45:06 UTC