Hugh, hello. On 2010 Nov 5, at 17:47, Hugh Glaser wrote:
> The issue I was pondering on was whether it was being suggested that a > server could avoid sending the Gigs of data in <http://example.com/about> > when asked for one of the many hash URIs in <http://example.com/about>, > such as <http://example.com/about#alice>, but just responding with the rdf > for <http://example.com/about#alice>. > I think that the hash is stripped off, as you say, long before the server > in any case; but even of it wasn't, then things would break because of > caching of <http://example.com/about>, etc.. > This was all in relation to whether hash URIs are a problem for big files, > but that has long gone in the stream of emails now, I guess :-) My understanding of Leo and Richard's document (and at this point, perhaps someone should ping them!) is that they recommend the "hash-URI" scheme (in their terms) "for rather small and stable sets of resources that evolve together. The ideal case are RDF Schema vocabularies and OWL ontologies, where the terms are often used together, and the number of terms is unlikely to grow out of control in the future." (see section 4.4, Choosing between 303 and Hash) I think that they recommend this precisely because it's an implication of this scheme that the entire underlying document has to be retrieved to learn about any Thing described in it, which is unappealing if the underlying document represents lots of data. I've just noticed that they describe the "slash-hash-URI" scheme in section 4.4, though without giving it a distinct name: "An example for a combination of 303 and Hash is: http://www.example.com/bob#this Bob, the person with a combined URI." Best wishes, Norman -- Norman Gray : http://nxg.me.uk
