See also: IRC log
approval of minutes
our next F2F is 18-20th April in Edinburgh, will send out logistics
planning for Wed-Thurs to be testing, Friday to be processing issues and the documents
pauld: please look to the Action Items list, it's out of hand!
yves: can we distinguish between
examples we exclude beacuse of barf dancing and those we
exclude because of broken examples?
... we also have some broken instance documents
pauld: please make sure you don't commit broken instance documents!
yves: need reviewing
<Yves> this one is easy to fix
<scribe> ACTION: gcowe to review and fix instance documents which fail validation [recorded in http://www.w3.org/2007/03/20-databinding-minutes.html#action01]
<trackbot> Created ACTION-121 - Review and fix instance documents which fail validation [on George Cowe - due 2007-03-27].
yves: telno has to begin with a '0' and yet it's an int
pauld: we started out making the element and attribute names human readable, but I think we've learnt that it's not clever to make an attribute name "telno", we should make it "integerAttribute" or some such to avoid semantics above the type
yves: looks like a toolkit
problem after all
... we need to do a re-run
pauld: wondering if it's still realistic to publish the report is ahead of the F2F
yves: having to do lots of
... needs more cosmetics
pauld: main concern is the accuracy
yves: we have tests marked as failed, need to be sure it's not my comparison
gcowe: on the failures - are you deciding by looking at the HTTP response, body?
yves: just comparing SOAP envelope
pauld: the perl captures the HTTP headers for extra info, but a standalone Java command would be better for those who don't understand the ways of CPAN
gcowe: working on such a thing
pauld: each toolkit has a
... contains per-toolkit info for the report
... and a list of examples to exclude, some are because of the examples are broken
gcowe: I'd like to test just basic or advanced patterns
yves: i've managed to do this in dashboard.java
gcowe: spent too much time on the barf dancing and excluding tests, now working on building the output.xml file from a Java command
pauld: would like to wrap this work up as a zip file so others can supply tests with their own toolkits
yves: george, you can edit the makefile to add your tests to the report
pauld: have a couple of tests up
my sleeve from non-java kits
... thanks for all your hard work, I need to catch up with it
pauld: nothing to report here,
... interesting news from IBM: http://www.alphaworks.ibm.com/tech/purexml
... looks like work around the databinding for DB2
... does anyone have DB2 experience/expertise?
gcowe: have some experience, but not using it currently
pauld: haven't quite understood it, yet, looks interesting for us
jonc: wondering about raising
more complex issues around this issue
... has some Java to test this, should we be looking to reuse this - as a post-processing step?
... views anyone?
gcowe: haven't yet looked at the java code you sent me
jonc: same issue applies to extension and element reference
pauld: valid issue, rasied at the original workshop, but how do we test it, do we need Java?
jonc: we could add this to the Java command which tests a schema
pauld: other tests we could add
for assertions, eg. UTF-8
... is this a pattern or an assertion?
jonc: I'd like to see this tested for as it's easily added almost by accident
pauld: note Philippe also added a useful appendix on the schema types and elements used within Basic and Advanced patterns
jonc: could be useful to cite the other examples
pauld: think we could write
something generically around the component model
... suggest adding the same constraint to the other examples
... is this for Basic or Advanced?
jonc: pathalogically cyclic
dependencies shouldn't be Basic
... e.g mandatory links
... with minOccurs="1"