|Ecore mapping discussion [message #577545]
||Wed, 10 September 2003 15:45
Originally posted by: ihc.sas-glas.mottmac.com|
I'd be interested in rationale and future plans in two areas where the xsd
ecore mapping loses information present in the XML Schema
1) mixed content (mixed = true)
Although I dislike the idea of mixed content, I don't have control over
the schemas that we will encounter. I sympathise with the the XSD approach
of simply having Object "contents". However, this mapping is not very
descriptive of the original XML Schema. Given that XSD 2.0 will have wider
schema-ecore mapping support I wonder if this implementation will
Only XML Schema enumerations based on NCName become ecore enumerations.
For enumerations based on xsd:string, the XML Schema enumeration info
appears to be lost in the ecore mapping. What's the rationale here - is it
because only NCName is guaranteed to be implementable as an enum in Java?
And as above, does the expansion of ecore mapping in XSD 2.0 mean that a
change in the mapping rules is possible here?
Powered by FUDForum
. Page generated in 0.18213 seconds