See also: IRC log
minutes 20th December approved
pauld: still to create editors' documents .. am working on them
yves: databinding often between
similar or the same languages
... restriction on ranges may be possible in some languages, such as ADA
... such constraints maybe validated using asserts even though the type may be looser
pauld: what are you hoping for here?
yves: I don't think it will be an issue, unless I hear more feedback
pauld: I guess this topic will crop up later
pauld: waiting for Vladislav
jonc: walks through his proposal:
... problems not just with schema patterns, but in the use of such patterns in Web services toolkits. Publishing in a WSDL allows people to test patterns against an endpoint
Ajith: what do you mean by consistantly named global element?
jonc: people may submit small
extracts of schema, but possibly larger schemas of which only
part of it is the feature
... we need an entry point to such schemas
... echo means the same schema structure is used for input and output
... we need an entry point for the echo method
pauld: we also need an instance document to test the echo operation
discussion of having more than one instance documents
pauld: instance documents useful when testing databinding toolkits using canned interactions, not so much toolkit to toolkit
<scribe> ACTION: jonc to raise a separate issue on submitting patterns without WSDL wrapper [recorded in http://www.w3.org/2006/01/10-databinding-minutes.html#action01]
<trackbot> Sorry, couldn't find user - jonc
pauld: awaiting text from paul ..
pauld: useful resource posted to our list:
pauld: has contacted Schema WG chair to reuse their list.
pauld: motivation was to get people to think about this imporant Schema pattern in common use, but known not to work with many tools. How to bucket such patterns: Basic, Advanced and do we document anti-patterns?
jonc: real world experience means
that this is an advanced pattern
... universally supported or well supported
pauld: do we qualify each pattern with a "star rating"?
<Yves> if this issue is about putting constraints on types (like integer), in that case it will be a null constraint, then it might not be an advanced pattern
jonc: sufficient to classify advanced or basic
pauld: personally uncomfortable about absence meaning not well supported for well known patterns
jonc: you'd like to demonstrate
completeness for such anti-patterns?
... we don't need to shoot for completeness of Schema. Just show things which give *good* experience
ajith: basic (v) advanced (v) antipatterns useful when generating a schema from a toolkit
pauld: and when humans author schemas
pauld:It seems like we're starting to agree that Basic is what works in practice and Advanced is more aspirational. We should decide if this is an advanced pattern or an anti-pattern next week.