RE: Rationale for restricted uses of "all"

> > It seems like I need a third content model group type:
> >
> > <all>
> >         <element name="A"/>
> >         <contiguous>
> >                 <all>
> >                         <element name="B"/>
> >                         <element name="C"/>
> >                 </all>
> >         </contiguous>
> > </all>
> 
> Wouldn't a nested xs:all do the job in this case, if that were
> allowed?

I see a orthogonality between ordering and contiguity. Could you make it
implicit that <all> implies contiguity? Yeah, sure... sequence has to by
definition; but <all> doesn't *necessarily* have to. Neither do occurrence
constraints necessarily have to imply contiguity in an <all> model context
(Gary Robertson first noted this). 

I could be nitpicking here, but isn't that what specs are all about? :0)

Received on Wednesday, 21 November 2001 15:52:18 UTC