- From: Jeremy Carroll <jjc@hplb.hpl.hp.com>
- Date: Thu, 11 Aug 2005 08:38:29 +0100
- To: Dan Connolly <connolly@w3.org>
- CC: Dan Brickley <danbri@w3.org>, public-rdf-in-xhtml-tf@w3.org
Point of this message: One possible solution to the bnode attribute problem is not to have one (or two) but to use URIs. This was the idea in marks x-pointer solution, that noone else liked. In essence the problem is we don't have appropriate URIs. I think it is arguable that we don't have appropriate URIs because, despite such a URI being in scope of 3986, there isn't any such scheme, which seems adequate motive for defining such a scheme. Discussion in two parts, first about the scheme, then about xhtml 2 =========================== The Scheme ---------- As part of my Jena work, I'm thinking of implementing RFC 3986/3987 (URI, IRI) and read the following text: [[ URIs have a global scope and are interpreted consistently regardless of context, though the result of that interpretation may be in relation to the end-user's context. For example, "http://localhost/" has the same interpretation for every user of that reference, even though the network interface corresponding to "localhost" may be different for each end-user: interpretation is independent of access. However, an action made on the basis of that reference will take place in relation to the end-user's context, which implies that an action intended to refer to a globally unique thing must use a URI that distinguishes that resource from all other things. URIs that identify in relation to the end-user's local context should only be used when the context itself is a defining aspect of the resource, such as when an on-line help manual refers to a file on the end- user's file system (e.g., "file:///etc/hosts"). ]] http://www.ietf.org/rfc/rfc3986.txt?number=3986 This got me thinking on whether a bnode identifier (with a slightly different syntax, e.g. local:name, where "local" is the URI scheme name and name is a string matching the path-rootless construct of RFC 3986), could be a URI in this sense. The key text in a definition of such scheme would be: [[ The local: URI scheme provides a method for constructing identifiers in a uniform way, consistent with other identifiers, for resources for which no other URI is known or appropriate, but which can be given an adequate interpretation in some end-user context. A "local:" URI is always interpreted in relation to the end-user context, and should not be used except where that context supports this scheme. This is similar to a file: URI without an authority component which should only be used where there is a contextual file system (the localhost); but differs in that the possibilities for the context of interpretation are less constrained. ]] RDF documents would be given as the example end user context, merging of multiple contexts could be discussed (the renaming function). This could be done easily by giving each of the contexts to be merged a local: URI and then interpreting the local: URIs within context A as relative references to be resolved in relation to the local: URI of context A. Context A: local URI for context: local:A/ local URIs in context A local:a local:b Context B: local URI for context: local:B/ local URIs in context B local:a local:b Merged context has local URIs local:A/a local:A/b local:B/a local:B/b Also, renaming within a context (i.e. the graph isomorphism of RDF) can be discussed: i.e. it is always possible to substitute all references for local:a by local:newa where local:newa is unused, and the overall meaning is unchanged). Hmmmm, by requiring renaming to leave the query and fragment parts unchanged, it would make the scheme usable for a wider range of applications. ========================= XHTML 2 ------- If we wanted to pursue this approach, we then go ahead with XHTML2 with no solution for bnodes, and separately start up a scheme registration and see how it goes. If the scheme registration fails (which of course it might) we then revisit for XHTML 2.1 ... It would be worth having a draft for such a scheme up pretty quickly so that it can be informationally referenced from XHTML 2 ... Jeremy
Received on Thursday, 11 August 2005 07:39:39 UTC