Re: rel=type or rel=profile, issue 92

On 20 Jan 2014, at 09:54, Wilde, Erik <Erik.Wilde@emc.com> wrote:

> last one, i promise...
> 
> On 2014-01-18, 00:38 , "Kingsley Idehen" <kidehen@openlinksw.com> wrote:
>>   Because RDF is supposed to enable you resolve these kinds of issues
>>   i.e., by way of vocabulary/ontology and the relation semantics
>>   expressed in the vocabulary/ontology.
> 
> well said. to me, the fundamental problem here is that outside of RDF,
> many things one the web are contextual. finding a certain link relation in
> one media type may mean something different from finding it in another.
> that's because each media type will often make the precise definition of
> what a hypermedia control means, found in the context of that media type.
> as long as the semantics fit the link relation roughly, all is well.
> 
> that's kind of against RDF's idea that things have context-free meaning,
> and this is why when reading the "profile" RFC (and many other link
> relation RFCs, i would guess), it gives you the semantic heebie-jeebies.
> most link relation registrations i am aware of very intentionally are
> fuzzy, trying to be a good fit for a certain (loose) class of assertions
> you might want to make, and leaving it to the context (the media type) to
> define what they precisely mean in some context.

The context we are looking at here is the HTTP header.

[[ http://tools.ietf.org/search/rfc5988#section-5
   The Link entity-header field provides a means for serialising one or
   more links in HTTP headers.  It is semantically equivalent to the
   <LINK> element in HTML, as well as the atom:link feed-level element
   in Atom 
]]


> 
> sadly, i cannot see a good way out of this, it's a bit of a different
> approach of how to define semantics. i simply wouldn't spend too much time
> with intense exegesis of RFCs; you are trying to read things into the spec
> that intentionally have been left out of the spec.

RDF allows very fuzzily defined relations too, just consider foaf:Person for
example:

[[
The Person class represents people. Something is a Person if it is a person. We don't nitpic about whether they're alive, dead, real, or imaginary.  The Person class is a sub-class of theAgent class, since all people are considered 'agents' in FOAF.
]]

But when I turn to understand what foaf:Person is about, I still refer to the http://xmlns.com/foaf/0.1/Person definition document.

I think looking at the RFC 6906 to know what "profile" means is not nitpicking. It has to be the starting point
of the investigation.

> 
> cheers,
> 
> dret.
> 
> 

Social Web Architect
http://bblfish.net/