WS Description WG telcon

23 Jun 2005

See also: IRC log


Allen Brookes, Rogue Wave Software
Roberto Chinnici, Sun Microsystems
Paul Downey, British Telecommunications
Youenn Fablet, Canon
Hugo Haas, W3C
Tom Jordahl, Macromedia
Anish Karmarkar, Oracle
Jacek Kopecky, DERI Innsbruck at the Leopold-Franzens-Universitšt Innsbruck, Austria
Amelia Lewis, TIBCO
Kevin Canyang Liu, SAP
Dale Moberg, Cyclone Commerce
Jean-Jacques Moreau, Canon
David Orchard, BEA Systems
Bijan Parsia, University of Maryland MIND Lab
Tony Rogers, Computer Associates
Arthur Ryman, IBM
Adi Sakala, IONA Technologies
Sanjiva Weerawarana, Invited Expert
Umit Yalcinalp, SAP
Rebecca Bergersen, IONA Technologies




<scribe> scribe: pauld

marsh: have to drop off around the hour, hugo to take the chair later

anish: has to drop off soon for WS-RX

minutes approved from last week's telcon

umit: driving to WS-RX

jjm: has to drop off early also

Action Item Review

?         2005-04-21: Pauld to craft, publish Common Schema structures
                      to WG for review for publication as WG Note, 
                      due 2005-06-28. 
?*        2005-05-12: Glen to add scoping example to primer, 
                      due eob 2005-06-16.
?*        2005-05-19: Umit to provide #none for Primer, 
                      due eob 2005-06-16. 
DONE [.3] 2005-05-31: Umit to incorporate these three points into new
                      text - 1) it's about the message, dammit, not the 
                      operation, 2) it's context-dependent, 3) for the 
                      contexts which we define as common, here are the 
                      things to be thinking about (unique GEDs, etc), 
                      due eob 2005-06-16. 
DONE [.8] 2005-06-01: Glen to formulate concrete async requirement 
                      for CG, due 2005-06-16.
DONE [.5] 2005-06-09: Jonathan to make a registration page for the 
                      July FTF, due 2005-06-16. 
?         2005-06-16: Amy to provide test cases for MEPs not described
                      in Part 2, due 2005-07. 
DONE [.6] 2005-06-16: Arthur to update Service Reference part of primer,
                      due 2005-06-23. 
DONE [.4] 2005-06-16: Arthur to fix composition models, due 2005-06-23. 
DONE [.7] 2005-06-16: Jacek to propose a non-XML example, probably 
                      MIME-based, for next telcon, due 2005-06-23.

[.1] http://www.w3.org/2002/ws/desc/#actions
[.2] http://www.w3.org/2002/ws/desc/4/lc-issues/actions.html
[.3] http://lists.w3.org/Archives/Public/www-ws-desc/2005Jun/0069.html
[.4] http://lists.w3.org/Archives/Member/w3c-ws-desc/2005Jun/0045.html
[.5] http://www.w3.org/2002/09/wbs/34041/wsdesc0705/
[.6] http://lists.w3.org/Archives/Member/w3c-ws-desc/2005Jun/0044.html
[.7] http://lists.w3.org/Archives/Public/www-ws-desc/2005Jun/0075.html
[.8] http://lists.w3.org/Archives/Public/www-ws-cg/2005Jun/0002.html

<hugo> Requirements from Glen: http://lists.w3.org/Archives/Public/www-ws-cg/2005Jun/0002.html



marsh: make sure you register for July f2f
... looking for host on West Coast in September (Bay Area)

<sanjiva> please make it second week of september ...

<sanjiva> not 1st week :(

+1 kids go back to school first week in september in the UK

<sanjiva> can we pick a date now? that's kinda consistent with the at least 60 days notice thing right?

daveo: mnot isn't active in the bay area, would other west coast area could be an option

marsh: thinking week of the 12th as a straw man

<sanjiva> +1 for the week of the 12th for me!

marsh: Kevin isn't here for Primer status review
... editing status, most of our documents are ready to go

Alternative Schema Languages

bijan: willing to publish as-is. two issues for extension writers regarding mixing and matching multiple type systems
... will send text

Jacek: wanted text regarding non-XML type systems. Suggested text turned out to be short suggesting use of mime-types

bijan: finds mime-type confusing example for alternative type systems. suggests java type system

umit: thought example was for primer, not alternate schema language note
... example seems to complex for primer, would prefer something else from mime-type

jacek: only alternative that came to mind was RDF/OWL which would be too complex, but likes bijan's suggestion of Java, exhibits import, etc
... may not have time over the next couple of weeks

marsh: do we really need this to declare victory? others outside the group could do this work and not impact our timeline

bijan: wonders about the value of our typesystem extensibility without examples
... but the value of doing the extra work is questionable

hugo: matter of priorities, we could drop this work to publish on time

<hugo> /me especially as it seems that there are details to iron out

hugo: sees value in the work, but can be done later


marsh: summary of Schema Workshop discussion
... i submitted a brief paper introducing LC124 and presented the issue to the Schema Working Group / Workshop

Workshop program, slides and papers: http://www.w3.org/2005/03/xml-schema-user-program.html

scribe: workshop saw value in our requirements
... schema group seemed unlikely to complain if we provided support for ignoring unknown content
... some members seemed enthusiastic about the notion of a schema describing what should be processed, not just validation

daveo: what was the reaction of where the annotation should take place

allen: Noah expressed some concerns about compatible evolution

pauld: people were interested in the notion this would better sell the PSVI as opposed to boolean validation

marsh: strong interest in the room for work in the area of versioning, but unlikely the schema WG would provide the solution

tom: so must-ignore seems to be a valid approach?

<kliu> jonathan, don't want to interrupt the conversation - just want to let you know that I am in the call and sorry for calling in late - got home too late last night

marsh: i was encouraged, there seem to be no obvious land mines

hugo: what impact will this have on our schedule?

marsh: seemed likely that this could be tractable
... had good feedback from Microsoft, that our products work in this way already

daveo: has suggested text, which I ran by Henry
... softens the relationship between the flag and the result, close to paul's original proposal but sharper


scribe: outlines the proposal
... there maybe other techniques beyond double schema validation which are also valid
... for removing unknowns

hugo: is there a default value for the flag?

daveo: will of the working group seems to be not to ignore unknowns

tom: do you think we can really get away with this?

daveo: this doesn't change how a schema processor works

marsh: few processors validate anyway
... one of the topics of discussion was relationship between data binding and validation, and they seem to be orthogonal

tom: seems to be true of at least Axis

daveo: and xmlbeans

pauld: xmlbeans is awesome, btw

marsh: sees this is this a policy assertion, rather than a change to schema processor
... was orignially concerned by this proposal, now encouraged

tom: we seem to have new information, and seems like a useful thing, was concerned about implementation

daveo: encouraged by how many implementations already do this already

marsh: any tricky aspects could be flushed out using CR

pauld: would love it to be the default, but probably a step too far for this working group

tom: happy to move forward

marsh: do we want to make this standard beahviour, and not even have a switch at all?

tom: wow, you're really trying to make paul happy :-)

daveo: would like this to be standard behaviour

pauld: would make our spec simpler

daveo: you might think this is a big change too late, and the WG prefers optionality

marsh: (speaking for Microsoft) seems like it's on by default in our products

pauld: there are brittle implementations out there, but we tell customers NOT to use those implementations

tom: we could change Axis to work in this way quite easily
... do we use a MUST for this

sanjiva: we don't talk about processors, so how do we specify this

marsh: we can find semantics for this

sanjiva: most implementations don't do validation

daveo: best position could be to make it optional, but the default

hugo: leaning towards this

<sanjiva> pauld: I also noted that this approach does not allow one to detect an error element vs an extension to be ignored

marsh: so proposal is to still have a switch

daveo: ws-security had a more strict model, maybe a must understand model is needed

sanjiva: without any validtion, hard to tell difference between an error and a new version

marsh: using henry's algorithm, UPA does assist here

daveo: java code provides extra implicit validation, e.g. mapping a string to a java int

sanjiva: maybe a more dynamic language might help here, and wouldn't be able to distinguish between an error and a new version

<sanjiva> :-)

marsh: ability to mark in WSDL that a service does strongly validate does seem to have value
... any objection to providing a flag?

non heard

marsh: has to drop off
... descision tree, flag, default, where it goes (schema, element, endpoint, description, etc)

pauld: we're loosing critical mass, and this is v.important

daveo: will crispen up our proposal

marsh: any objection to ending the call now?
... but we still have some other work to do, but many people going to WS-RX

meeting adjourned

Summary of Action Items

[End of minutes]

Minutes formatted by David Booth's scribe.perl version 1.126 (CVS log)
$Date: 2005/06/23 16:20:26 $