- From: Tim Berners-Lee <timbl@w3.org>
- Date: Sat, 3 Nov 2007 00:03:10 -0400
- To: cdr <_@whats-your.name>
- Cc: Semantic Web <semantic-web@w3.org>
On 2007-11 -02, at 13:28, cdr wrote: > >> __$ curl -I http://www.w3.org/2007/ont/unit.rdf >> Link: meta;rel=meta > > i had a question about how to link to metadata for an Image or > other binary file, from the header, and i think this addresses it > > Link: http://abs/path/to/meta.rdf;rel=http://purl.org/dc/elements/ > 1.1/description > > ^- correct use with absolute URIs? > > will tabulator follow anything here. or only specific predicates > denoting availafixble metadata? It will follow any URI for the file, but it must be rel=meta -- not a URI at all. It currently processes it in just the same way as an HTML link element. This gives the HTML WG control of a bit of HTTP spec, effeciively. We could argue that 'rel=meta' here is a relative URI relative to a common default. > > can i just throw all the triples (eg EXIF) right into a single link > header using commas to seperate? including strings? Not in this implemention. But I must admit thinking of HTTP headers Prefix: dc=http://purl.org/dc/elements/1.1/ N3: <> dc:title "Another interesting picture". The prefixes would have to affect all the N3: headers, ie be deemed to have been given first, as http is unordered. Tim > >> Tabulator implements this on the cleint side, and pulls in the >> metadata. >
Received on Saturday, 3 November 2007 04:03:25 UTC