2006/7/2, Jan Algermissen <[EMAIL PROTECTED]>:
Without any intent to argue about this, here is a thought though:
IMHO, the media entry acts as a negotiation point (for the metadata
and the actual media resource)
+1
and this suggests using the redirect (again, IMO).
No, this means they are different representations of the very same
resource, and suggests that content negotiation could take place at a
given URI.
If you also want separate URIs for each representation( which is IMO a
Good Thing), then a Content-Location should be present in any
negotiated response. Of course, the "specific" URI of a representation
must not in turn use content negotiation.
So you have three URIs:
- the URI of the "entry", using content negotiation to return either
of the following:
- the URI of the Atom Entry Document (application/atom+xml)
representation ("media link entry")
- the URI of the "other" representation ("media resource")
For example, using a widely deployed approach for content negotiation:
- "entry", with con-neg: http://example.net/entries/PictureOfTheBean
- "Atom" representation: http://example.net/entries/PictureOfTheBean.atom
- "non-Atom" representation: http://example.net/entries/PictureOfTheBean.jpg
I've almost finished my PaceUnifiedEntryEditing, which I explained
about 10 days ago:
http://www.imc.org/atom-protocol/mail-archive/msg05441.html
I'd be glad if you wanted to help me putting the final touches on it.
--
Thomas Broyer
