Re: [GRAPH] graph deadlock?

On 22 Dec 2011, at 11:56, Pierre-Antoine Champin wrote:

> On 12/22/2011 12:59 AM, Steve Harris wrote:
>> On 21 Dec 2011, at 08:54, Antoine Zimmermann wrote:
>> 
>>> Le 20/12/2011 21:55, Steve Harris a écrit :
>>>> [skip]
>>>> 
>>>> I think most(?) people agree that a URI should denote/name/something
>>>> a graph, or some other entity, but not both at the same time. The
>>>> problem is that people don't follow this rule in RDF now*, don't
>>>> follow it in quads as implemented now, and I don't think they will
>>>> follow it in the future.
>>>> 
>>>> So, does that break RDF, or does it break their applications?
>>>> 
>>>> If it just breaks people's applications, then we can write what we
>>>> would like to happen in the document, and people who do the Right
>>>> Thing™ will be fine, and people who don't will suffer in some way.
>>>> 
>>>> If on the other hand it breaks RDF, it's probably already too late,
>>>> and we have a problem.
>>>> 
>>>> - Steve
>>>> 
>>>> * e.g. http://blog.iandavis.com/2010/11/04/is-303-really-necessary/
>>> 
>>> I haven't read the whole post but where do you see a URI which is used to denote two different things at the same time? How do you know it denotes 2 things simultaneously?
>> 
>> Well, if I have a document like:
>> 
>> <http://example.com/foo> a <Thing> .
>> 
>> and then I dereference http://example.com/foo, and get a 200 and a document back, isn't http://example.com/foo both an instance and a document?
> 
> Ian's point was not to advocate that the same URI identifies both a
> thing and a document. Read the added section ("Update Nov 5") where
> '.../toucan' identifies an animal, and '.../toucan.rdf' identifies its
> description in RDF/XML.
> 
> His position was more about the necessity of 303-redirect for URIs
> denoting "non-documents". His suggestion was to respond with a "200 OK",
> but using the "Content-Location" header to indicate the indirection.

Ah, sorry, I missed that - apologies to Ian for misrepresenting his position! I just quickly googled for some relevant terms, looking for indications that people were doing it in the wild, and it seemed like the blog post was on the subject.

FWIW I agree with him that a 303 is a very high cost to pay.

> Of course, this contradicts the TAG's position that "200 OK" can only be
> used for so-called "information resources". But that does not implies
> that URIs denote several things.

No, sure.

- Steve

-- 
Steve Harris, CTO, Garlik Limited
1-3 Halford Road, Richmond, TW10 6AW, UK
+44 20 8439 8203  http://www.garlik.com/
Registered in England and Wales 535 7233 VAT # 849 0517 11
Registered office: Thames House, Portsmouth Road, Esher, Surrey, KT10 9AD

Received on Thursday, 22 December 2011 13:38:15 UTC