[XML-DEV Mailing List Archive Home] [By Thread] [By Date] [Recent Entries] [Reply To This Message]

RE: RDF Interpretation of XML documents (was Re: [xml-d

• To: "'Thomas B. Passin'" <tpassin@c...>, xml-dev@l...
• Subject: RE: RDF Interpretation of XML documents (was Re: [xml-d
• From: "Bullard, Claude L (Len)" <clbullar@i...>
• Date: Wed, 21 Aug 2002 09:20:10 -0500

So we need a way to construct the schema for the 
application domain user but still be able to add 
information for the missing labels?   

Old SGML habits lead me to not leave a mixed-content type 
in there anyway.  We've had <text></text> elements forever.  

As for John's semanticless container nodes, 
those are a problem.  The fight on the X3D list over 
these was intense because they are a way to create 
a context-free parse for those who want that, 
typically object-oriented programmers who say that 
fields should map to attributes and fields can 
contain nodes, which of course, breaks XML, to 
which they responded, then XML is flawed.

And then there is the convenience of OR groups.

len

-----Original Message-----
From: Thomas B. Passin [mailto:tpassin@c...]

Yes, this is just how I look at it.  Actually, you can represent any xml
document as either a node-centric graph (elements become nodes) or an
arc-centered one (elements become arcs).  Even mixed-content documents can
be so represented if you think of imaginary "text" elements that contain the
character data.  In the first, the nodes are labeled and the arc labels have
to be inferred; in the second it is the reverse.

Either way the derived graph is incomplete in its KR content (labels are
missing), and this simply reflects the fact that xml is about syntax, not
semantics.  The semantics would supply the missing labels.

• Prev by Date: Re: namespace reprise
• Next by Date: Re: namespace reprise
• Previous by thread: Re: RDF Interpretation of XML documents (was Re: [xml-d
• Next by thread: RE: RDF Interpretation of XML documents (was Re: [xml-d
• Index(es):
  • Date
  • Thread