norman.walsh.name: Comments on /2003/12/12/housekeepinghttp://norman.walsh.name/2003/12/12/housekeeping/comments.atom2012-02-13T09:07:58.038169ZComment 1 on /2003/12/12/housekeepinghttp://norman.walsh.name/2010/09/25/oauth#comment00012003-12-14T20:47:30Z2003-12-14T20:47:30ZJohn Clarkda39a3ee5e6b4b0d3255bfef95601890afd80709

I'd like to ask for a feed covering all of the "Technology" topic - I'm very interested in many of the different subtopics, which I think are nicely encapsulated by the parent topic.

Comment 2 on /2003/12/12/housekeepinghttp://norman.walsh.name/2010/09/25/oauth#comment00022003-12-15T15:24:51Z2003-12-15T15:24:51ZEd Daviesda39a3ee5e6b4b0d3255bfef95601890afd80709

I think TBL & DanC's comment is absolutely correct. However, the interesting question is: did they suggest any other mechanism to get to the metadata?

I've just read the Architecture of the World Wide Web, First Edition and can find no clues there. I wonder if maybe an HTTP GET-META message or something should be created.

Comment 3 on /2003/12/12/housekeepinghttp://norman.walsh.name/2010/09/25/oauth#comment00032003-12-18T05:23:53Z2003-12-18T05:23:53ZDavid Menendezda39a3ee5e6b4b0d3255bfef95601890afd80709

I agree with Dan C. and Tim B.L. that a description of an essay is a different resource from the essay itself, but I don't think it follows that you therefore can't do content negotiation.

If these essays had (say) French translations, you would want to provide URIs for (1) the essay, (2) its English version, and (3) the French version. A request for (1) would return the same material as (2) or (3) depending on the language preferences of the server and client. The server indicates that (2) or (3) is being returned through the Content-Location header.

GET always returns a representation of a resource. If someone tries to GET this essay and says they want RDF/XML, why not send them the metadata? The Content-Location header makes it clear that the data being returned has its own identity in addition to representing the essay in some manner, and the data itself would probably describe that relation.