Henry Story has rather a provocative post with REST without RDF is only half as bad as SOAP, in which he criticises domain-specific XML formats, in particular that used in the Atom Publishing Protocol.
Although I agree with most of Henry's arguments, I've also banged my head against the wall with several of them in various forums a lot of times (granted, as has Henry). So I don't personally expect anything like general adoption of the RDF approach to vocabularies around XML in the foreseeable future. But that certainly isn't a showstopper for the Semantic Web, or even necessarily a cause for pessimism.
[that pic's from 07/2003, wow, he looks quite different now - must do an updated photo set]
A significant proportion of the significant bits of information carried by virtually every XML-based language can be expressed in RDF. XML is also relatively easy to transform into other formats, such as those used for RDF serialisations.To put it another way (as I'm sure someone else has already), when looking at angle brackets, elements and attributes, all I see now is triples, triples, triples...
See also: micromodelsÂ@en