Re: AW: Can hash URI description lookups be made to scale?

Patrick Stickler writes:

> On Jan 30, 2004, at 13:28, ext Leo Sauermann wrote:
> 
>> the hash (#) had also caused me some trouble, I woiuld avoid it and
>> use it only in namespaces.
> 
> I would advise avoiding it nearly everywhere. Terms in vocabularies
> are first-class resources no less important than web pages, service
> portals, image files, etc. And we want to have just as robust and
> useful URIs for those resources as for any resource.
> 
> "Just say NO" to fragment identifiers!
> 
> ;-)

Hash URIs do have their disadvantages (such as, there's no good way to
get them to work with an MGET verb in HTTP), but they are nice in
lower-tech areas.

For example, say I'm working on a small RDF vocabulary and I put an
ontology describing it at <http://example.org/vocab>. If it defines
terms of the form <http://example.org/vocab#whatever>, then anyone
trying to learn more about the terms can simply dereference them to get
their definitions. If they had been of the form
<http://example.org/vocab/whatever>, then the site would need (1) lots
of ontology files, each describing one term, or (2) redirections from
the various URIs to the ontology URI.

This isn't an issue on more sophisticated sites (e.g., anything running
URIQA), and it isn't practical for super-large vocabularies (e.g.,
WordNet), but hash URIs have their place.
-- 
David Menendez <zednenem@psualum.com> <http://www.eyrie.org/~zednenem/>

Received on Monday, 2 February 2004 01:17:53 UTC