- From: Jeremy Carroll <jjc@hplb.hpl.hp.com>
- Date: Thu, 21 Oct 2004 16:46:39 +0100
- To: "'public-rdf-in-xhtml task force''" <public-rdf-in-xhtml-tf@w3.org>
I have an outstanding action to clarify my concerns about literals. Unfortunately I still haven't allocated long enough to do the job properly, so I am afraid I'll sketch the worries without adequately linking to the doc ... forgive me if I make silly mistakes ... Starting point: Literal text in the html doc gets translated into RDF XML Literals, except where an explicit datatype is mentioned. (Motivation: text in html can have markup included) Observations: 1) There does not seem to be any method for adding a plain literal to the metadata. e.g. if I have a triple of the form <http://example.org/doc> dc:creator "Jeremy Carroll" . where "Jeremy Carroll" is a plain literal then I cannot include this in an XHTML 2.0 document. 2) Language is a problem: Suppose I have the following triple in an XHTML document <http://example.org/doc> dc:creator "A Guide to Markup"^^rdf:XMLLiteral . and suppose that somewhere there is an in-scope xml:lang="en" attribute within the XHTML document. Since rdf:XMLLiteral is a datatype, it does *not* have the capability to be language modified. The RDF Core WG, when making this decision, noted that such language markup could be added into the XHTML e.g. instead of the above triple, we have: <http://example.org/doc> dc:creator "<span xmlns=\".../xhtml2...\" xml:lang=\"en\">A Guide to Markup</span>"^^rdf:XMLLiteral . I suggest that this insertion of <span> with xml:lang is always done. 3) I think it will surprise some RDFers to find all their metadata coming out as rdf:XMLLiterals or even (if 1 is addressed) any metadata that is not declared as not rdf:XMLLiteral. I don't see anyway round this, and personally hope this will encourage the RDF community to take embedded XHTML within RDF more seriously. e.g. providing tool support for searching within XHTML as XHTML within RDF content. I think that's about it. I hope there is enough detail there to be intelligible. Jeremy
Received on Thursday, 21 October 2004 15:47:27 UTC