> Yes, that's right, WXS is very wrong to mix both aspects.
We agree on this.
> Who knows which is the best
> choice for the author of my original RNG schema? No automatic tool can
> be sure to make the right choice and, if it needs to ask the question
> for any such choices (this one is only an example), it might become as
> difficult and complex for the user to answer those choices than to use
> WXS :-( ...
I think there is a right choice here, and that is not to try to abuse the
global/local element distinction for something for which it was not
designed.
>> That's not how I interpret the spec. Which schema processors disallow
>> this?
>
> My interpretation is based on a discussion with Henry Thompson:
>
> http://lists.w3.org/Archives/Public/xmlschema-dev/2001Nov/0188.html
>
> And the spec seems rather clear:
>
> http://www.w3.org/TR/xmlschema-1/#mg-props-correct
I don't agree. See the discussion at
http://lists.xml.org/archives/xml-dev/200202/msg00935.html
Fortunately the major schema processors I've tried all agree with me (XSV,
MSV, SQC, MSXML). Are there any that adopt your interpretation?
James
