Mukul Gandhi scripsit:
> Following is an use case, that probably makes the notion of "optional"
> in XML documents *necessary*,
>
> Let's assume, that a book object may be represented in an XML format
[snip]
> i.e, an ISBN number of book may be represented either as an attribute
> or as a child element of element book, but not at both of these
> places. To define a schema for this need (at least with XSD I
> believe), a schema author I think *must* make the attribute "isbn" and
> the element "isbn" as optional and constrain a singular cardinality of
> the semantic information "isbn" via another schema constraint.
That is indeed an XSD 1.0 limitation. It does not apply to RNG, where
representing a choice between an element and an attribute is trivial.
I don't know if it applies to XSD 1.1 or not. It does not apply to
Schematron either, where everything that is not forbidden is permitted.
There is a more general point: it's not always possible in any schema
language to represent the exact specification, and applications may need
to do further validation of the input.
--
LEAR: Dost thou call me fool, boy? John Cowan
FOOL: All thy other titles http://www.ccil.org/~cowan
thou hast given away:
That thou wast born with.
