Re: URI aliases for RDF terms?

On Wed, 2011-11-09 at 17:50 +0000, Richard Cyganiak wrote:
> On 2 Nov 2011, at 16:05, Eric Prud'hommeaux wrote:
> >> Either way I'm not a fan. Yes, the rdf:type URI is a bit ugly, but so is foaf:name. It's too late to change it IMHO.
> > 
> > +1
> > Excruciating precision is a cost of doing business in an unambiguous domain.
> 
> <http://n.w3.org/rdf/type> is no less precise than <http://www.w3.org/1999/02/22-rdf-syntax-ns#type>, but has significantly lower cost.
> 
> W3C has maneuvered itself into a dead-end by assigning uselessly verbose URIs to really important things. This is a roadblock for uptake in syntaxes where long URIs are expensive.

My gut reaction here is to agree.   There are some details, though....

> IMO W3C should:
> 
> 1. Assign short URIs in a *single* namespace for all RDF and RDFS (and possibly OWL and XSD) concepts *now*

Do we know of any principled way to make namespace boundaries?

How do we decide if OWL gets combined?  How about rif or geo or org?

Some ideas:

1.  like software modules -- a collection of strongly-connected
elements, small enough to be understood in its entirety (hopefully
without too much effort).    This would rule out including OWL in with
RDF.  One could argue whether RDF and RDFS belong in the same one.   (I
would argue that they should be combined.)

2.  by the group that developed it -- this would combine RDF and RDFS,
but keep it separate from XSD, OWL, RIF, etc. 

3.  by the group that has ultimate legal control over it -- this would
combine RDF, RDFS, OWL, XSD, RIF, and might or might not include org at
this point.   (looking at the license at http://www.w3.org/ns/org I'm
thinking W3C does control it.)   I'd suggest good URLs if we go this
route would be like "http://id.w3.org/type".

I see schema.org going for the 3rd option, with the justification that
it's easiest for users.

Given the current tools, they might be right, but I can't help thinking
that with better tools, (1) would be superior.

> 2. Leave it up to individual WGs to adopt those short URIs at their own leisure
> 3. Leave it up to implementers to add support for them already
> 4. The RDF WG should *not* do anything about them in RDF 1.1, but perhaps in RDF 2.0

Not even mention them?  Not give any advice as to how they should be
used?

I, for one, would *love* to see the developer community start to
seriously do sameAs reasoning in every RDF data consumer, so folks could
freely use aliases.   It seems like it has to happen sooner or later to
make this thing (the Linked Data ecosystem) work.    

I'd hate to have a big one-time pain cutting over RDF URIs at the cost
of general machinery, when that machinery is going to be needed again
and again as we go forward.

   -- Sandro


> Best,
> Richard
>

Received on Wednesday, 9 November 2011 18:25:14 UTC