W3C

XML Schema Patterns for Databinding Working Group Teleconference

05 Feb 2008

See also: IRC log

Attendees

Present
Jon Calladine (BT)
George Cowe (Origo Services Limited)
Paul Downey (BT)
Yves Lafon (W3C)
Regrets
Chair
pauld
Scribe
pauld

Contents


Administrivia

pauld: sorry, didn't get an agenda out for today!

pauld: but it's the standing items, plus moving Basic Last call as discussed at last week's F2F.

pauld: minutes from 2008-01-30, 2008-01-31, 2008-02-01 F2F (W3C ERCIM Sophia Antipolis, France)

2007-12-18 teleconference

2007-10-30 teleconference

2007-10-09, 2007-10-10, 2007-10-11 Interoperability Workshop (BT, London, England)

2007-08-01, 2007-08-02, 2007-08-03 F2F (Origo, Edinburgh, Scotland) approved!

ISSUE-2: test suite

george: generated reports for 10 toolkits, took longer than I expected as new patterns caused barfing

<Yves> http://www.w3.org/2002/ws/databinding/edcopy/report/all.html

george: don't have the environments for BEA, ZSI etc

pauld: OK, we can't carry on doing this for ever

http://www.w3.org/2002/ws/databinding/edcopy/report/basic.html

we've a lot of white, can we be confident that means we didn't run the tests, not that they barfed?

george: new examples means we get an unbalanced view if we don't rerun a toolkit, we need to remove it

pauld: prefer to highlight where it barfed rather than when we didn't run it

maybe I could regenerate this: http://www.w3.org/2002/ws/databinding/edcopy/venn/venn.html

yves: I have the list of which examples aren't good

http://lists.w3.org/Archives/Public/public-xsd-databinding/2008Jan/att-0003/process.jpg

yves: I can add "exclude error" which will turn white to red for barfing

george: concerned that some tools have more green than others, just because we haven't rerun them

pauld: seems ok but we need to change the text "(White) indicates the test was not run. Often this is because the toolkit was unable to consume the XML Schema or WSDL document." for skipped in the report

pauld: I will fix that so as far as testing goes, we need Yves fix, my venn regenerated/improved and change to text

pauld: also nice to see if we can't get Adrian to help re-run. Will do this once we're "done".

<gcowe> Spring (Castor) and BEA Weblogic testsuite not successfully run yet.

<gcowe> Axis 1.4, Axi2 1.3 (ADB), Axis2 1.3 (XMLBeans), .Net, Mono, Spring (JAXB2), Spring (XMLBeans), Spring (XMLBeans), XFire and XFire (XMLBeans) all rerun successfully. Can Yves rerun report?

yves: more vendor tools, such as BEA, the better

<gcowe> and PHP too! nearly forgot about that one.

<Yves> the report is up to date

pauld: would be good to tidy the report up ahead of a director's call

<Yves> I will rerun after editing the toolkit.xml

ISSUE-12: identifying a conformant schema

yves: the service works with the directly with the patternsdetector.xsl latest uses the standard W3C XSLT service

yves: if you want to run on a local schema, you can always download the command line version

george: you need to work from the command line, not as simple as using the web site

pauld: wonder if there is a quick win using an online service, sounds like a similar issue to the microformats stylesheet/plugins

pauld: who knows, the stylesheet may find its way into Oxygen, XMLSpy, etc

we need to do more work on this and the collection before we can progress Advanced

pauld: suspect we need more work, and as we add more patterns, we'll need to do more work to make them coherent

george: I spotted a couple with maxOccurs which seemed unnecessary

Basic Last Call

pauld: thanks Jon for your patterns text. Will try to incorporate the text in a human readable, but precise way (a challenge!)

jonc: I think we need more explanation of how and why we use XPaths

pauld: worried we sail down a slippery slope, and don't want to introduce chatty or informal text into a specification, but will try to make it readable.

pauld: Any objections on moving Basic to second Last Call, expecting it'll go to PR>

none heard

yves: four weeks is the minimum, seems sufficient for me. We didn't make many changes.

pauld: we made a resolution to move dateTime to advanced based on testing, but how do we feel about that now?

george: happy for it to be in advanced, does have issues.

yves: they're ok, so long as they're between 1970 and 2038 and timezone qualified, and 5 digit years!

pauld: so we're still OK with Advanced? tools map it to a programming dateTime type which isn't a good fit.

pauld: OK, so Last Call of one month, including non-normative patterns text.

our next call is in two weeks time, when we'll look at Advanced

Summary of Action Items

[End of minutes]

Minutes formatted by David Booth's scribe.perl version 1.133 (CVS log)
$Date: 2008/02/19 17:31:45 $