W3C home > Mailing lists > Public > public-xg-webid@w3.org > December 2011

Re: a remark on the webid spec

From: Henry Story <henry.story@bblfish.net>
Date: Wed, 14 Dec 2011 09:38:44 +0100
Cc: public-xg-webid XG <public-xg-webid@w3.org>, Pierre-Antoine Champin <pierre-antoine.champin@liris.cnrs.fr>
Message-Id: <594F8CC9-24E6-443A-9CA4-192738D06F64@bblfish.net>
To: Mischa Tuffield <mischa@mmt.me.uk>

On 14 Dec 2011, at 02:33, Mischa Tuffield wrote:

> Hello, 
> 
> I thought I would have a look at the RDF and SPARQL in the spec too.  And I have a few comments too. These are but nit picking, will start with the RDF. 
> 
> The three fragments of RDF could all be interpreted to be slightly different. 
> 
> 1. In the RDF/XML and the Turtle all of the foaf:Person IRIs are set explicitly to : https://bob.example/profile. But this is not the case in the RDFa, there is it dependant on the location of the html document. I would prefer all of them to be baseIRI dependant (like the RDFa). Regardless of the fixed adopted i think consistency is the way forward. 
> 
> You could either 
> 	a) change in the Turtle s;/bob:me/<#me>/ and from the bob @prefix 
> or
> 	b) change the RDFa to be s/"#me"/"https://bob.example/profile#me"/
> 
> b) is probably an easier fix, but a) seems neater. 

Interesting point. I would also be in favour of a) i.e., for relative URIs, it makes the point better.
Ah.

On the other hand there is the issue that people may want to publish

profile.html
profile.rdf

and have both of those be served up at 

profile

with content negotiation.

In that case you need the profile.html and profile.rdf to be talking about

profile#me

and there you need to have the relative uri be <profile#me> . 
We do I think point to the "how to publish rdf" document, but I don't think that explains that,
 so perhaps that is an advanced feature. Perhaps we should have a little section on that

"Publishing multiple representations of a WebID"

Does someone wish to write that up?


> 
> 2. In the RDFa the foaf:name "bob" inherits a language tag. I don't like language tagged names, they don't feel right. And plus there no language tag in the Turtle or the RDF/XML. 
> 
> You could add a xml:lang="" to the foaf:name span

yes, you'd need language tags that span very large numbers of regions. 
+1

> 
> 3. Turtle consists of 9 triples, rdf/xml of 8, and the rdfa has 7 triples. 
> 
> Missing triples are basically foaf:weblog from the rdf/xml, and foaf:weblog along with rdf:label

Yes. I think it is good to have the consistency with the diagram.


>> [snip, answered separately]
> 
> _____________________________
> Mischa Tuffield PhD
> http://mmt.me.uk/
> http://mmt.me.uk/foaf.rdf#mischa
> 

Social Web Architect
http://bblfish.net/
Received on Wednesday, 14 December 2011 08:44:46 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Wednesday, 14 December 2011 08:44:46 GMT