Re: [soapbuilders] practical guide for common denominator of schema contstructs?
- On 26 Apr 2006, at 00:49, timastaJ wrote:
> Does anyone have any pointers to documentation/best practices thatThe W3C XML Schema Patterns for Databinding Working Group is
> capture the lowest common denominator of schema construct support in
> major platforms? i.e., it is colloquially known that xsd:union does
> not have wide spread support in J2EE app servers, but has this been
> formally captured anywhere (besides per-platform READMEs). Has anyone
> taken a stab at producing this laundry list? I'm not looking for a
> i rant here... ;)
currently trying to address this task in its "Basic Patterns"
document. The result will be a W3C Recommendation and a validation
tool which will indicate if a given XML Schema or WSDL is likely
to give "a good user experience with databinding tools".
We're also working on an "Advanced Patterns" document which
are a set of XML Schema patterns in common use which should
be better supported by databinding tools (but typically aren't).
This work is progressing slowly, thanks in no small part to
a lack of participation. In particular we've seen very little
interest from vendors, many of whom still claim to support
"the whole of schema" when they evidently don't, or
"the latest version of foobar works with xs:union" when the
current version of foobar has been in the wild for 3 or so
years and plainly doesn't.
The Working Group is conducting technical discussions in public,
and you are all welcome to contribute patterns and experiences
Identifying the aspects of schema which doesn't work with
tools hasn't been the approach taken so far, but I continue
to wonder if "naming and shaming" tools which purport
to abstract away XML using an XML Schema as an IDL and yet
fail to map 'xs:choice', etc be a more fruitful activity.
- On 4/26/06, Paul Downey <paul.downey@...> wrote:
> Identifying the aspects of schema which doesn't work withgo for it :)
> tools hasn't been the approach taken so far, but I continue
> to wonder if "naming and shaming" tools which purport
> to abstract away XML using an XML Schema as an IDL and yet
> fail to map 'xs:choice', etc be a more fruitful activity.