::scr semantic

Ben scr@thegestalt.org
Mon, 5 Aug 2002 13:02:36 +0100


On Mon, Aug 05, 2002 at 12:21:40PM +0100, jo walsh wrote:
> 
> > But the semantic web built from the hideous cruft that is XML and all
> > its bastard offspring, is, I'm certain, most definately not what Ted has
> > been looking for.
> 
> rdf != xml 
> RDF/XML is one form of rdf serialisation.
> n-triples is another, n3 is a third: http://logicerror.com/n-triples/
> 
> xml looks useful. techwars don't.
> oops, i said i'd stop, didnt i.

I'd like to know why you think XML is a better representation format for
these type of relations, rather than, say, S-expressions.

Particularly since the trust metric part of Rivest's SDSI (which appears to
me to be a much sounder basis for a semantic architecture than either anything
derived from an SGML source, which would also need a separate tamper-proofing
and trust architecture anyway) is already based on a cut-down form of 
S-expressions. [0]

Remember folks, XML parsing, in and of itself is actually an expensive operation.

Techwars are, IMHO, very useful if they prevent good people from wasting vast amounts
of time on what are the wrong technologies. (X509 global certificate hierarchies
and global namespaces anyone?)

Ben
[0] SDSI-style S-expressions have a trivial transformation into a limited form of XML
anyway, for the people who like XML-on-toast.