Databinding WG Teleconference

6 Mar 2007


See also: IRC log


Jon Calladine (BT)
George Cowe (Origo Services Limited)
Paul Downey (BT)
Yves Lafon (W3C)



minutes from 30th Jan, 6th Feb, and the F2F approved

gcowe: we suggested F2F 18-20 April

pauld: heard a lot of love for the WG at the Workshop last week
... challenge is to turn that into participation
... maybe we could have an event to promote our work with implementers

jonc: a day at a F2F, maybe May-June to promote our testsuite sounds good

gcowe: are we ready for that, we need to do more work to be in a position to run an event

pauld: a F2F to prepare our testsuite and documentation might be a good way forward
... I'd like to get the likes of XFire and WSO2 onto a test report and have them compete to be interoperble

jonc: would that encourage others to participate?
... a publicised webinar might get people in and not involve commitment on their part

pauld: a screencast might be an idea, something can run in their own home
... show checking of a schema using the patterns detection service with a toolkit.

yves: publishing a version of the test report will help a lot - people will send comments

Test Suite

jonc: been looking for missing patterns in BT WSDLs
... choice still being throw up as unknown
... in particular choice with different cardinalities

<pauld:> wonders why "we're" using choice, given it's not a basic pattern

pauld: are you making the patterns more generic, or adding more patterns?
... just tred carefully making the patterns too generic

gcowe: need to look into this further

pauld: best to build up with fine grained patterns

Test Report

yves: been working on the checking of the response XML to request XML
... attributeDefault required a specific check
... close to having something "good enough"
... ready to work on aggregation
... verification is based on the example name, e.g integer triggers integer comparison

pauld: working on pattern detection for schemas in the wild

Running the Tests

pauld: toolkit.xml excludes examples which cause barfing
... runit HTTP POSTs instance documents and builds output.xml
... as usual I'm the weakest link, and you guys are making me look worse, so that's good news :)

gcowe: so the one service does all the operations?

pauld: yes

gcowe: barf-dancing for Axis took an hour and a half with the deploy/undeploy step

Last Call Comments

pauld: still not updated EXIT, sorry!
... have had some new comments
... philippe asked "is choice a basic pattern"
... difficult to tell at present, philippe has offered to help build a table at the end of the spec for the use of each schema element
... Hugo also commented on the complexity of XPaths, I'll try and tidy them up, and write an introduction

jonc: we've found that some XPaths have to be complex
... we've found the XPath reported for "unknowns" isn't always valid

gcowe: it looks like an XPath, but maybe the stylesheet is breaking the lines up and making them invalid?

pauld: sounds like a bug! I wrote that code, maybe you could send mail and I'll fix it

jonc: a certain toolkit "voldermort" is rejecting our schemas, which contain circular references
... can we write an XPath to detect that?

pauld: possibly within a single schema document
... or a WSDL
... but not with import and include, not easily that is
... we have the same issue with inheritance
... we could apply an assertion in our basic spec against circular dependencies
... we need to add assertions to the validation service -for encoding etc

jonc: would circular dependencies and mandatory links is asking for trouble
... minOccurs="n>0" is asking for trouble

pauld: raise an issue

OK we'll meet again in two weeks time


Summary of Action Items

[End of minutes]

Minutes formatted by David Booth's scribe.perl version 1.128 (CVS log)
$Date: 2007/03/20 18:06:33 $