- From: Jason Cupp <jcupp@esri.com>
- Date: Fri, 19 Nov 2004 23:47:30 -0800
- To: "'public-esw-thes@w3.org '" <public-esw-thes@w3.org>
With or without a frag id, the server is getting the entire URL line, and can do whatever it wants. Frag IDs in practice have been used by HTML clients, but it's not a rule that frag IDs can only be acted on by the client. RFC2396 says: >The semantics of a fragment identifier is a property of the data resulting from a retrieval action... Therefore, the format and interpretation of fragment identifiers is dependent on the media type [RFC2046] of the retrieval result. It's nailed down for HTML clients, but for RDF user-agents we don't have to stick to the HTML model. RFC2396 continues: > Individual media types may define additional restrictions or structure within the fragment for specifying different types of "partial views" that can be identified within that media type. Who is to say that GETing http://foo/#term1 for the RDF media-type, has to return an RDF/XML file pre-existing on the server? If I have an RDF service running, I'm not going to base it operationally on files, HTML style, but on an RDBMS. Requesting http://foo/#term1, might return a "partial view" in RDF/XML consisting of the resource http://foo/#term1 and it's outward-bound properties, a sub-graph. Either way, getting an actual RDF/XML file or getting RDF/XML as serialized by a service, the client is going to interpret the result, which in RDF's case, could mean checking to see if http://foo/#term1 is actually in the return graph, in a subject position at least. Even if GETing RDF does follow the HTML model, we shouldn't be afraid of parsing large documents. If I have an RDF application, which aggregates from other RDF sources locatable via HTTP GET. It doesn't make sense to dump from your RDF catalog all the assertions from a source you wish to get updates from, then re-assert those again. You compute a digest hash from the new GET, compare it with what you fetched before, then only if the hash is different do you go on with the expensive operation of cosuming the updated RDF/XML. - Jason
Received on Saturday, 20 November 2004 07:48:03 UTC