> If you say
> the whole points is gone... and if you do content-negotiation
> on top of
> -that- URI (which you perfectly could), it's getting even
> more confusing
> since you are requesting a JPG image and the server returns
> an SVG image
> depending on your browser capabilities.
> > Mime-headers & stuff only happen post request.
> Yes, that's the beauty of it.
It seems the two of you are talking about two things:
- how to tell what the content type of one thing is
- how to ask for the content type of many things
The content type of one resource is clearly best described via
the Content-Type header in the HTTP response.
The content types of many things is ambiguous right now. But with a little
bit of content negotiation it is pretty straightforward.
You can ask for a resource (like a collection of things) and specify
the return type (via the 'Accept' header). This is what happens with
http://machine/something/ (Accept: text/html)
If you used Accept: text/xml you could return an XML list of the resources,
meta-data like Content-Type.
But then it wouldn't be 'standard' or common. This is what WebDAV addresses.
A way to use HTTP
to get a clear description of the resources (either collections or not) in a
rather then a presentation description.
Alternatively, you could still use the HTML output & add an attribute to the
line items for the content type: