See also: IRC log
pauld: next week is July 4th, does anyone object to meeting?
pauld: OK, we meet next week as usual :-)
pauld:if there's interest we could book a table at The Rex cinema monday of our F2F:
planning to publish Last Call Working Draft around next F2F
would like to not formally meet during August
planning to work towards an interop event for advanced patterns, possibly in Boston during December close to XML 2006
pauld: presented to WS-I, had some push back, which I blogged: http://blog.whatfettle.com/archives/000397.html
discussion around using Relax mapping as a profile
jonc: little experience or interest in the WG around this approach
pauld: suggest we ensure Trang generated patterns are covered in our Advanced document
pauld: done nothing on this since
we last spoke
... suggest task will become easier once we have test cases published
pauld: took exploding sed scripts
from Yves and turned themround into one stylesheet per target:
instance, soap11, soap12, wsdl, schemas
... being driven by a single patterns.xml
yves: suggest using xinclude from xmlspec to include patterns and examples
pauld: separated out examples from patterns, associated by links
pauld: been testing with
toolkits, building a list of what we need to know to test a
.. been working with .NET, Mono, Axis 1.x and SOAP4r toolkits
<scribe> ACTION: ylafon to investigate use of wiki [recorded in http://www.w3.org/2006/06/27-databinding-minutes.html#action01]
<trackbot> Created ACTION-50 - Investigate use of wiki [on Yves Lafon - due 2006-07-04].
pauld: alternative is to get people to send mail and just build an index from that
still an open issue in BP, not over yet, WG will solict schema WG's advice
jonc: published a pattern, the
Pete Hendry pointed an obvious issue with the order of the
... we need to either have more examples, or have separate patterns for specific combinations
... xsi:type may also play a role here
... which way do we go?
... with instance documents, should we use xsi:type to test failure conditions
pauld: is the order of the
filters a design consideration?
... we can also have more than two simple types, combinations explosive
jonc: design consideration could
spell out what makes sense, could duplicate xml schema
... been using tools to play with combinations, but how exhaustive should I be?
... maybe we should do some inductive reasoning, just try 2, and 4 or 5
pauld: so I'm wondering what we expect tools to do with union "xs:date xs:string"
jonc: should we also include anti-pattern instance documents?
pauld: see no problem in having
'bad' instance documents
... just annotate bad documents is important
discussion of 47! combinations, may be reduced by non-sensible combinations