Re: [httpRange-14]: New Draft Finding "Dereferencing HTTP URIs"

I (Ed Davies) wrote:
>> Would it be a fair summary of this finding to say that
>> a 303 response to an attempt to access someURI with a
>> location field set to someOtherURI and the assertion:
>>
>> <someURI>
>>   <http://www.w3.org/2000/01/rdf-schema#seeAlso>
>>     <someOtherURI> .
>>
>> are equivalent?
>> ...

Leo Sauermann replied:
> That is semantically true, but not precise enough. seeAlso links can be 
> used to link across any RDF documents on the web that describe the same 
> resource. The relation between <someURI> and the 303 redirect to 
> <someOtherURI> (which would then be a html or rdf description of 
> <someURI>) is more specialized than seeAlso. Because the site admin of 
> someURI can only redirect to ONE such someOtherURI, the someOtherURI has 
> a special meaning of being "I, the domain holder, think the best 
> description of <someURI>".

Good points: 1) 303 can point to a non-RDF document whereas
rdf:seeAlso shouldn't, 2) 303 should point to an authoritative
document rather than just any old document which happens to say
something about the original resource and 3) 303 is functional
whereas seeAlso isn't.

> We could raise seeAlso into this more specific relation, but I would 
> rather go with (what also Dan Brickley has brought up in [2])
> foaf:isPrimaryTopicOf. Or my own pimo:occurrenceRef, or SKOS' 
> http://www.w3.org/2004/02/skos/core#isPrimarySubjectOf .

Aren't foaf:isPrimaryTopicOf and the like the wrong way round?
That is, they say that some resource O talks mostly about some
other resource S when what 303 says is that some resource S
is defined by another resource O.  For example, the following
makes sense:

<http://www.w3.org/People/Berners-Lee/card#i>
   foaf:isPrimaryTopicOf
     <http://en.wikipedia.org/wiki/Tim_berners-lee> .

but even if it wasn't a hash URI you wouldn't want to 303
from TimBL's personal URI to the Wikipedia article (but
rather to his own FOAF file).  isPrimaryTopicOf is functional
from the document to the subject whereas 303 is, as you
point out, functional from the subject to the document.

So, an equivalent of 303 would be something like
xxx:isDefinedBy or perhaps xxx:isDefinitiveDescriptionOf,
depending on your taste in property directions.

A resource which is a definitive description of something
need not have that thing as its (one and only) primary
topic.  It could define lots of things.  xxx:isDefinedBy
would be a functional property whereas foaf:isPrimaryTopicOf
is inverse functional.

> And I would hope to raise isPrimaryTopicOf/isPrimarySubjectOf as an 
> extension into the RDFS or RDF vocab, as it is rather important across 
> many applications on top of rdf. (otherwise, foaf and skos and many 
> heretics like me need to fight it out).

I think that makes sense from the point of view of
not having to declare too many namespaces.  Apart from
that, there doesn't seem to be any harm in sharing -
it's rather the point, isn't it?

Ed Davies.

Received on Monday, 4 June 2007 16:29:23 UTC