Re: [xml-dev] Venetian Blinds vs Garden of Eden patterns for industrysta

[Mukul Gandhi <gandhi.mukul@gmail.com>]

On Thu, Oct 28, 2010 at 5:22 PM, David <dlee@calldei.com> wrote:
> The one thing I'm most confused about is what the namespace issue is.
> Why do the different patterns impose different namespace issues ? Arnt all
> the elements in the schema in the same namespace, and reguardless of if you
> implmeent the XSD elements as local or global types, they are in the same
> namespace, and instance documents need follow the same rules ?
>
> What does where global/vs/local element declaration change wrt namespace use
> in the instance doc ?

I think specifying namespace constraints in XML schema gets better in
1.1 version where we can specify targetNamespace indicator on
xs:element and xs:attribute instructions (not available in 1.0), which
allows us to have one schema document validate an XML document which
may have arbitrary composition of namespaced element and attribute
instances.

With XSD 1.0, to specify heterogeneous namespaced element and
attribute declarations in an schema, we have to specify xs:import in
the parent schema document, and specify XSD element/attribute ref= at
desired places to allow different namespaced elements/attributes than
the targetNamespace on xs:schema instruction. This required us to use
as many schema documents (via imports) as different namespaces XML
document was composed of (i guess that was painful as number of
different namespaces grew in an validation episode). In XSD 1.1, we
can have only one schema document which can handle any number of
namespaces, via localized targetNamespace instructions (though the XSD
1.0 facility of xs:import is still available in XSD 1.1).




-- 
Regards,
Mukul Gandhi