- From: Mark Birbeck <mark.birbeck@x-port.net>
- Date: Thu, 17 Jan 2008 21:36:17 +0000
- To: "Ben Adida" <ben@adida.net>
- Cc: RDFa <public-rdf-in-xhtml-tf@w3.org>
Hi Ben, > That said, I would suggest that we not add this specific syntax to our > test suite: I'm worried that some parsers would then add that <#base()> > to the URIs in their triples (which would not be entirely without reason.) I wasn't proposing we add this to anything. :) I was just pointing out that what you were talking about "fits nicely" with things I've been working on. However, you've made me realise that there is an interesting issue that we probably need to draw parser authors' attention to. If you take the 'current URL' from the DOM, you may get something like this: <http://www.example.org/group.html#ben> However, the base URI for parsing should be this: <http://www.example.org/group.html> and all triples need to have that as the default subject. This seems right to me intuitively _anyway_, since how you navigate to a document should not alter the triples that are generated about that document. But interestingly enough, it looks like we get this 'for free' because we set @about="". This is because the algorithm for resolving relative paths will set the fragment identifier of the resulting URI to the fragment identifier from the relative path that is being converted. The full algorithm is here: <http://gbiv.com/protocols/uri/rfc/rfc3986.html#relative-transform> We probably need to add a point of clarification about this, as well as some tests in the test-suite. Regards, Mark -- Mark Birbeck, formsPlayer mark.birbeck@formsPlayer.com | +44 (0) 20 7689 9232 http://www.formsPlayer.com | http://internet-apps.blogspot.com standards. innovation.
Received on Thursday, 17 January 2008 21:36:25 UTC