|
Architecture |
XML | RDF
In October 1997, I demonstrated to the RDF Model and Syntax Working Group that if somebody wants to use RDF in a valid XML document, they can:
By way of reviewing the recent RDF draft [members-only; @@update to public draft], I constructed a conforming SGML document consisting of the following files (attached):
- test.rdf
- the examples from the spec (transcribed by cut/paste) plus a prologue with element/attribute declarations for elements such as DC:Creator used in the examples; The prologue refers to...
- rdf.dtd
- a DTD fragment that is a transcription of the grammar in the rdf spec. It allows the RDF: namespace qualifier to be redefined by the document from which it is referenced by the normal SGML parameter entity rules. (Some might call this overengineering, but I believe it is a correct literal interpretation of the RDF spec)
- rdf.dcl
- an SGML declaration that I frankenstiened together from the HTML 2.0 spec and the XML spec, since nsgmls burped on the one from the XML spec and I wasn't in the mood to figure out why
I don't think rdf.dcl expresses the recent case-sensitivity decision.
I congratulate the RDF spec editors: the examples all parse correctly without modification. Output of nsgmls is attached.
I present this to the w3c-xml-sig as evidence that the layman namespace proposal[members only @@update to Jan '99 REC] is completely compatible with SGML '86.
I have not attempted to validate the document with an XML processor (I don't have one installed in a way that's convenient to operate like my nsgmls setup). I invite others to do so and report thier findings.
Fyi...
My nsgmls installation reports:
connolly@beach ../rdf9710[1206] nsgmls -v I: nsgmls version `0.4'