Re: Is 303 really necessary - demo

On Fri, Nov 5, 2010 at 1:53 PM, Jörn Hees <j_hees@cs.uni-kl.de> wrote:
> If I GET http://iandavis.com/2010/303/toucan i retrieve a document (I'll call
> this A) with rdf statements.

This is not correct. You receive a response with an entity: the
representation. (Here entity is used in the rfc2616 sense)

> If I GET http://iandavis.com/2010/303/toucan.rdf i retrieve another document
> (I'll call this B), which in this case happens to have the same content as A,
> but could be different, can't it?

I could return a different entity, but I wouldn't recommend it. You
might want to if you link to multiple descriptions of the resource.

>
> Now: how can I say that I don't like A without saying that I don't like
> <http://iandavis.com/2010/303/toucan> ?
>
> If your answer is going to be "say you don't like B" again, please explain
> what happens if A and B don't have the same content.
>

How do you currently refer to the entity transmitted in a HTTP
response? You don't - they have no names. How do you say you are
offended by something written on twitter's home page when the entity
it sends changes every second.

> Is there some magic involved saying that any ?s with a
> ?s <http://vocab.org/desc/schema/description> ?d .
> is not a document but a real-world object?

No. But the description document and the entity returned from the
request to /toucan says it's a dbp:Toucan. I would put more credence
in explicit statements than implicit ones.

>
> Or is there some magic involved that if toucan and toucan.rdf give you the
> same content that one of them is a real-world object then?
>

No.


> If not, how can I find out that <http://iandavis.com/2010/303/toucan> is one
> and A is only one of its descriptions?
>

Look at the data - it states it clearly.

> Jörn
>
> PS: is there a summary of this discussion somewhere?
>
>

I'm afraid not, it's only been going a few hours. I haven't seen
anything that fundamentally challenges the idea yet, i.e. something
that would make me rewrite it. I am seeing several responses arguing
that it confuses the thing with the document but I explicitly show how
it doesn't in the blog post so I think that comes from people's
default assumptions. There are some responses saying "don't do that -
use fragments instead", but that's no help for the millions of
resources already deployed with slash uris and the many people who
prefer that style. Other responses have been like yours, seeking more
clarity on the ideas.

Ian

Received on Friday, 5 November 2010 14:42:52 UTC