Re: XML vs. RDF ==> see infoset domains


> I was in a meeting the other day where TimBL was trying to explain the
> difference between RDF and XML.  He said something vaguely like "XML
> only lets you say things about parts of documents; RDF lets you say
> things about anything."

An XML document is nothing more than a stream of unicode characters. The XML
1.0 recommendation contains a proper set of EBNF productions which define
the _abstract syntax_ of XML. The XML Infoset is nothing more than a subset
of this abstract syntax, the subset which has been declared relevent (i.e.
it is not relevent whether an attribute value is contained within a pair of
single or double quotes, it is not relevent what the order of attributes is,
it is relevent what the order of elements is.

There are no semantics attached to XML _itself_ rather specifications attach
semantics to a particular XML grammar e.g. RDF/XML, Topic Maps, perhaps some
XML representation of KIF etc.

Certainly one can represent the XML Infoset in an RDF syntax (the RDF XML
surface syntax, or an RDF abstract syntax (e.g. a set of triples)). It would
be the RDF MT that attaches a semantics onto this RDF Infoset

In this case, it is correct that the _RDF rrepresentation of an XML Infoset_
only says something about the particular XML document, but so what?

I would counter with an obvious answer: an XML representation of KIF, or
perhaps a non-RDF representation of DAML/OWL would allow one to say anything
that the underlying KIF or OWL MT allows, perhaps something about anything.


Received on Saturday, 2 February 2002 11:53:12 UTC