Re: Semantic content negotiation (was Re: expectations of vocabulary)

>> (a) most URIs are not dereferenceable, because common rules
>> for dereferencing them do not exist, or they are designed not to be;
>
> What do you mean "common rules" for derefernecing?  What is the  
> rule to
> dereference an HTTP URI, like http://www.w3.org?  Can you de- 
> reference it?

One relies on the protocol to define a method for dereferencing. For  
http: and https:, that is defined, and I shan't belabour the  
mechanism. For other protocols, that is not the case.

>> (b) most HTTP URIs do not dereference to a useful RDF representation;
>
> So what?

You tell me: I quote:
---
Xiaoshu:
"""
The agent takes this statement should further dereference the
http://bar.com/newfoaf#knowsWell and probably returns back a  
statement like,
http://bar.com/newfoaf#knowsWell rdfs:subPropertyOf foaf:knows .
It is not at http://eg.com/foo where the inference is done.
"""

Reto:
"""
Sorry, that's nonsense. Not only property URI's are not necessarily
dereferenceable and the possibly available graph representation may or
may not contain that statement - do you know about any FOAF client
behaving as you're suggesting it should? I don't, and I know I wouldn't
want to install it on my mobile phone of limited resources.
"""

Xiaoshu:
"""
Are you sure you are talking RDF? The only thing that is not  
dereferencable
is literal values because they are not URI.  But literal can only be an
object, not subject and property.
"""

>> (c) using fragment identifiers gives you a slightly better
>> chance of dereferencing to a useful representation, because
>> of HTTP's mechanics.
>
> How so? How will http://foo.com/#abc give me a better chance to  
> dereference
> than http://foo.com/abc?  Where did you get this idea?

Try it; hopefully the reason why will occur to you.

$ wget http://www.holygoat.co.uk/owl/2005/05/photo/ConcreteDocument
...
11:35:40 ERROR 404: Not Found.

$ wget http://www.holygoat.co.uk/foaf.rdf#RichardNewman
--11:36:28--  http://www.holygoat.co.uk/foaf.rdf
            => `foaf.rdf'
Resolving www.holygoat.co.uk... 65.98.65.147
Connecting to www.holygoat.co.uk|65.98.65.147|:80... connected.
HTTP request sent, awaiting response... 200 OK
Length: 48,726 (48K) [application/rdf+xml]

100%[====================================>] 48,726        54.44K/s

11:36:29 (54.34 KB/s) - `foaf.rdf' saved [48726/48726]

Using fragment identifiers makes the ontology the resource and the  
terms fragments of it, which gives you better odds of getting some  
RDF back, given a document-oriented HTTP server like Apache. Usually  
the resource will be the ontology URI.

Using slashes coins a new resource for each term, which means there  
is unlikely to be RDF available unless you explicitly request the  
ontology URI.

>> Reliance on the dereferenceability of a term, even an HTTP
>> URI, is foolhardy.
>
> You haven't truly comprehend the essense of semantic web yet.  Read  
> this
> info: http://esw.w3.org/topic/FollowLinksForMoreInformation

Xiaoshu, I do this for a living.

URIs are opaque identifiers. That is all you can rely on.

Received on Sunday, 30 July 2006 18:41:48 UTC