Hi,
Frans Englich wrote:
Hello,
In an XML format of mine I need embedded XHTML, information for human reading,
documenting the "object" the document instance describes. I have hesitations
on how to do that in the best way.
Currently I do like this:
<xsd:import namespace="http://www.w3.org/1999/xhtml";
schemaLocation="http://www.w3.org/2002/08/xhtml/xhtml1-strict.xsd"/>
<xsd:element name="test" type="xhtml:Block"/>
However, from what I can tell, the content of the <test> element isn't XHTML
any longer, but just a brick of the document I built, labeled as what the
targetNamespace says. I see namespaces as "identifiers" for XML applications,
and any 3rd party, such as a a XSLT sheet, no longer sees XHTML, but must
learn my particular format.
That was what I _first_ thought, but then I realized that:
<test>
<div></div>
</test>
didn't validate without <div> being in the XHTML namespace. Apparently, it
"is" still XHTML.
So I'm confused. What if I /didn't/ want it to be XHTML but be my format, and
only borrow the XHTML complexType as a building block? (perhaps it's a weird
unrealistic question)
I interpret it as that WXS:targetNamespace doesn't matter(not that I mind in
this case :) ); from whatever namespace a building block emerges from, is
what they'll have.
I used libxml2 2.6.16 for validation.
The mechanism you describe matches the machanism for 'included'
schemata. If you want to 'borrow' the components for XHTML, you need
create a totally different XSD: copy the XSD for XHTML, remove the
specified targetNamespace or better set it to the targetNamespace of
your main schema + hope that wildcards, if existing, will still make
sense. Or just copy & paste the needed pieces into your main schema.
The targetNamespace has its purpose; changing it, creates components
which are totally different; just their names and structures are equal,
which could theoretically happen with any XSD by _accident_.
Greetings,
Kasimier
|
