Jeni's Musings

(it feels like there should be a “this is the real resource” value for the rel attribute of <link> or <a>, but I don’t know of one).

Hmm, I suppose “alternate”’s not what you want? I guess you mean something like Atom’s “self”.

which gets assigned the URL

http://www.example.com/queries/4328

I’d canonicalise and then hash the query parameters to get the URI. That way, a particular query always leads to the same query URI, but at the same time an observer cannot infer whether anyone has sent the same query before.

The tricky problem with any query-as-resource setup is: do you garbage-collect stored queries eventually? If yes, when?

• If you keep them around for too long and there are too many possible queries, you open yourself up to a DoS attack from someone who keeps generating new queries in order to deplete the server’s resources.

• If you don’t keep them around for very long then the URIs generated are useful only in the moment because they soon turn into 404s. As IDs in RDF they would be useless.

On balance, this leads me to conclude that the query-as-resource approach is probably only feasible for authenticated requests.