Re: Associating EARL with a page

Hi Charles

I must say that your discussion about connecting EARL and Annotea
intrigued me so much, that I used the weekend for coding a Python/
RDFlib based Annotea server, to learn more about Annotea and RDF. 

PyNotea is Open Source, and released under the W3C license, and can
be found here: http://www.u-moe.org/PyNotea/. I have tested it 
with AnnoZilla.

I will use PyNotea as a playground. Maybe annotations can be useful for
manual accessibility checking?

Regards,
Nils Ulltveit-Moe


søn, 10,.04.2005 kl. 17.20 +1000, skrev Charles McCathieNevile:
> Hi folks,
> 
> this is clearly a topic we need to think about. It's actually pretty mch a  
> general RDF question, not specific to EARL.
> 
> There are a few ways that we can connect EARL to the page it is about.  
> Most of the obvious ones are, I think, not very good because they involve  
> relying on the author of a page, and I think in many many use cases where  
> people will want to find EARL it will be whenthe author hasn't created or  
> linked to it.
> 
> This is the reason why I would not go very heavily down the path of using  
> HTML's link with something like rel="meta", nor a URIQA server (an  
> extended HTTP server that will store and return metadata for any resource  
> as special methods), although both of those are reasonable options for  
> authors to use and are not bad as such.
> 
> It's also one reason I don't like the idea of a "well-known-location"  
> convention, such as storing EARL in an /earl directory or something. I  
> also dislike that for other reasons and think it is a bad idea.
> 
> For the rest of us third party evaluators, it would be helpful to have  
> somewhere to put EARL and for the rest of us EARL consumers it would be  
> nice to have somewhere to look for it. Eric Prud'hommeaux said at the face  
> to face that his new annotea server should be able to handle EARL quite  
> happily - it's a server you configure that receives and serves RDF. There  
> are various other RDF servers around, too. And the nice thing is that soon  
> if you send any of them a query written in SPARQL, you should find out  
> what they know.
> 
> I actually have an informal action item to follow this up with EricP -  
> apparently he needs to see an example query to configure his server, so  
> here is an attempt at one for the current spec that asks for anything  
> about an assertion where the assertion has 6 basic properties (The  
> OPTIONAL stuff at the end is the attempt to gather all the other info. It  
> is the part I am least sure of):
> 
> PREFIX earl: <http://www.w3.org/WAI/ER/EARL/nmg-strawman#>
> PREFIX dc:   <http://purl.org/dc/elements/1.1/>
> 
> SELECT ?subject, ?test, ?result, ?assertor, ?date, ?mode, ?some, ?more,  
> ?spare, ?stuff
> 
> WHERE
>    ( earl:Assertion earl:assertedBy ?assertor )
>    ( earl:Assertion earl:result ?result )
>    ( earl:Assertion earl:subject ?subject )
>    ( earl:Assertion dc:date ?date )
>    ( earl:Assertion earl:mode ?mode )
>    ( earl:Assertion earl:testCase ?test )
> OPTIONAL {
>    ( earl:Assertion ?spare ?stuff )
>    OPTIONAL
>      ( ?stuff ?some ?more )
> }
> 
> If I have the query right, this will pretty much tell you anything known  
> about any Assertion that has some value for each of the 6 basic properties  
> I have noted. There are no constraints - I will start a seperate thread on  
> querying EARL, because I think this should be noted in the technical spec  
> and explained in the tutorial one...
> 
> cheers
> 
> Chaals
> 
-- 
Nils Ulltveit-Moe <nils@u-moe.no>

Received on Sunday, 10 April 2005 19:34:41 UTC