WS Description WG telcon -- 4 Jan 2007

<Jonathan> ACTION: Jonathan to make sure parameters are added to the URI in the HTTP binding even when no whttp:location appears. [recorded in http://www.w3.org/2007/01/04-ws-desc-minutes.html#action01]

<scribe> Scribe: Roberto

Approval of minutes

Minutes approved

Action items

[Interop]
?         2006-11-30: [interop] John Kaputin to create a test case 
                      with "required=false". 
?         2006-12-14: [interop] Jonathan to fix transferCodings - 
                      add control group

[WG]
?         2006-09-21: Jonathan to check periodically that SPARQL has 
                      added schemaLocation.
?         2006-12-14: plh to come up with a more detailed proposal for 
                      CR112 if possible
DONE [.5] 2006-12-21: Jonathan to reference the Versioning document in 
                      the WSDL Primer.
DONE [.3] 2006-12-21: Jonathan to respond to Ashok Malhotra/WS-Policy 
                      on WSDL 1.1 component indicators.
DONE [.4] 2006-12-21: Jonathan to raise new issue adding 4 new assertions 
                      as per Arthur's note.


Current Editorial Action Items

Note: Editorial AIs associated with LC issues recorded at [.2].

[.1] http://www.w3.org/2002/ws/desc/#actions
[.2] http://www.w3.org/2002/ws/desc/5/cr-issues/actions_owner.html
[.3] http://lists.w3.org/Archives/Public/www-ws-desc/2006Dec/0086.html
[.4] http://lists.w3.org/Archives/Public/www-ws-desc/2006Dec/0085.html
[.5] http://lists.w3.org/Archives/Public/www-ws-desc/2006Dec/0094.html

Administrivia

Jonathan: WG extended six months
... from the implementor's call, there was no progress on the logging issue
... no visible progress on dashboard
... it will take ~2 wks to fix the logging issues and get testing done between WSO2 and Canon
... 19 issues still open
... at the pace we're at, it should take ~ 3wks to get the issue count to zero

for PR we also need to update the documents

Jonathan: for PR we also need to update the documents
... still a few resolutions not incorporated
... another area: validator progress
... we got coverage of 50% of the assertions
... it shouldn't block us from moving forward to PR
... it could take a month to increase coverage

Arthur: do we have dates for these events?

Jonathan: no, extrapolating from current pace

Arthur: would like to see a schedule

Jonathan: would like to set up a time to test each other's endpoints in the week of Jan 21st
... looking for a 4-6 hour time window
... if we can do that by the end of January, it's going to take about 2 weeks to get to the Director's call
... it would go to the AC in two weeks, then the AC gets 4 weeks
... so ~10 weeks from the hand-off to REC
... so we are looking for Recommendation at the end of April
... may need to add a few weeks to address any issues, rewrite test cases, etc.
... e.g. issues with gzip transfer coding
... would like to send LC comments through the working group
... only substantial comment is (4) in http://lists.w3.org/Archives/Public/www-ws-desc/2007Jan/0011.html
... does anyone else plan on reading the document and send comments? deadline is next Friday

Paul: people can comment outside of the WG too

Jonathan: my comments were not WSDL specific, really
... in WSDL we make special provisions for multiple inline schemas
... would like such a schema not to be flagged as violating any patterns

<scribe> ACTION: Paul to report back on which test cases in the WSDL test suite fail the basic patterns, with suggestions on how to address the issues [recorded in http://www.w3.org/2007/01/04-ws-desc-minutes.html#action02]

Jonathan: Policy last call comments by Charlton

http://lists.w3.org/Archives/Public/www-ws-desc/2006Dec/0055.html

Charlton: main issues are policy intersection and security considerations
... also "policy expression (4)" is not editorial

Asir: how many comments are WSDL-related?

Charlton: 3
... "XML Element Attachment" (a note), already filed issues for two other ones

Jonathan: potential issue in the item under "Policy Expression (4)"

Charlton: theoretical interest in defining an abstract policy for improved sharing
... analogy with interface/class relationship

Asir: policy doesn't deal with assertions

Charlton: placeholders for assertions
... offers to write up some examples

Monica: idea from BPEL TC; it's a template with placeholders

Jonathan: it looks like a general policy issue, not a policy one
... any other substantial issues we can look at today?

Charlton: no

<scribe> ACTION: Charlton to come up with a list of editorial comments and WSDL substantive issues [recorded in http://www.w3.org/2007/01/04-ws-desc-minutes.html#action03]

Philippe: it's still useful to let them know that there is interest from the group in an issue

<monica> +1

MTOM description

Jonathan: any news?

<asir> Jonathan, I am signing off

Philippe: XMLP WG meets on Wednesdays

Jonathan: postpone the issue till next week

CR117

Youenn: issue with special characters in parameters
... hard to parse the URI in that case
... possible solution escape these characters or just urlencode all parameter values

Philippe: currently we don't do any escaping
... last time we discussed this, we decided not to do anything about it

Jonathan: is it a duplicate issue?

Philippe: duplicate of CR053

<plh> http://lists.w3.org/Archives/Public/www-ws-desc/2006Jul/att-0001/20060629-ws-desc-minutes.html#item06

Philippe: url escaping happens at the transport layer

Arthur: you can prevent "/"s by defining your own type

Jonathan: or use NMToken

Arthur: even if you encode a slash, when the URL is decoded how do you reconstruct the parameter values?
... we'd have to define a particular escaping

Jonathan: unclear when to encode things, except to prevent illegal characters

Arthur: encoding won't by itself fix the problem
... are we going to have a primer example of constraining the schema type?

Jonathan: # could make the URL invalid too
... 2 #s seem to be invalid

Tony: are there any security issues?

<plh> http://lists.w3.org/Archives/Public/www-ws-desc/2006Sep/0034

Philippe: nothing beyond what the URL spec already handles

<plh> "No percent-encoding mechanism, as

<plh> defined in section 2.1 of RFC 3986, is performed on the replacement

<plh> value. "

<pauld> notes this is a common issue, Microsoft Outlook Webmail constructs URIs from the subject line, which generates URIs you can't dereference with mails sent with '+' and '&' in the title

Jonathan: restricting schema types is a partial solution, won't work well for book titles with whitespaces

Philippe: whitespaces will get encoded on the wire
... but there can be information loss

Jonathan: don't want to force users to encode data themselves
... should we mandate or allow an escaping mechanism? what about precluding slashes (which we won't be escaped in any case)?

Arthur: no need to exclude anything, it can be done via a schema type

Jonathan: "The%20Bible" in user data would be bad

Arthur: distinction between what's legal in an URI and how URIs are transported

Jonathan: xs:anyURI allows more characters than the URL spec does

Arthur: even in ASCII, to parse a URL you have to know about the special characters

Jonathan: are spaces even an issue?

Arthur: some research needed

Jonathan: with slash, couldn't I have a full URI in the data and the template contain just one parameter?
... http:location="{foo}"

<plh> [[ As a consequence, if

<plh> this IRI is an absolute IRI, the {address} property of the Endpoint

<plh> component is ignored. ]]

Jonathan: need to clarify if the absolutization of the URI precedes or follows parameter substitution
... the results can be different in the two cases (e.g. with .. in the parameter value)
... multiple issues
... if we allow slash characters, the resulting URLs may be ambiguous
... preference is not to outlaw this case

Arthur: we could require the serialization from XML to URI to be invertible

Jonathan: don't want to mandate that
... e.g. the case of existing services which I'd like to describe
... it's the use cases of WSDL as a general HTTP description language

Arthur: we could define a more constrained URI style
... leave the URI style as-is and define a new style which guarantees invertibility

Jonathan: it seems hard to define such a style in general terms

Arthur: we can do it in a fairly direct (but constraining) way

Youenn: it may be too complex

<scribe> ACTION: Jonathan to analyze CR117 further [recorded in http://www.w3.org/2007/01/04-ws-desc-minutes.html#action04]

WS Description WG telcon

4 Jan 2007

Attendees

Contents

Approval of minutes

Action items

Administrivia

MTOM description

CR117

Summary of Action Items