Re: Linking back to sameas.org?

Hugh, Toby

Thanks for the follow-up on this, and well, I'm pretty much convinced 
now that rdfs:seeAlso is the simplest way to achieve the backlink to 
sameas.org
And thinking twice, I'm now also convinced  that neither adding any 
specific semantics to this link nor specifying the class of sameas.org 
URIs would provide anything more than rdfs:seeAlso standard 
interpretation : go there to find more, I trust them.

So I think in the next release of lingvoj.org I will replace in each 
description the bunch of local sameas links by a single rdfs:seeAlso 
sameas.org/foo. I guess it would not be difficult for dbpedia folks to 
add it also to dbpedia URIs? Actually, supposing sameas.org becomes a 
sort of institution :-) , RDF search/navigation tools could built-in the 
sameas.org service extension as an option, so that publishers do not 
even need to include such links anymore. I can imagine a little "sameas" 
tab on Tabulator menu.


Bernard

> On 08/06/2009 20:59, "Toby A Inkster" <tai@g5n.co.uk> wrote:
>
>   
>> On 8 Jun 2009, at 12:22, Bernard Vatant wrote:
>>
>>     
>>> http://sameas.org/html?uri=http://www.lingvoj.org/lang/fr provides
>>> 16 equivalent URIs (including the original one).
>>> At http://www.lingvoj.org/lang/fr I've gathered painfully only 10
>>> of those :-)
>>> But now that sameas.org is alive, why should I care maintaining
>>> those sameAs links locally?
>>>       
>> I imagine that sameas.org found many of those links at lingvoj.org.
>>     
> You imagine right.
> Or at least to start with, as I recall that was one of the first sites that
> started this particular hobby, when I realised that we had generated a bunch
> of language URIs ourselves, and I wanted to link to the others.
>
> In fact Bernard raises an issue dear to my heart, which is that publishers
> of Linked Data should be able to do just that; and that others can be
> facilitated in providing and maintaining the links, in particular without
> putting any load on the data publisher to be troubled with any of it.
> These sort of separations are good engineering practice. I have long liked
> the idea that the linking is knowledge of a separate sort to the substantive
> content. It certainly often has different temporal characteristics.
>
> So I think what Bernard means is that he needs reliable service(s) that will
> look after his hard-won links, and allow him to maintain the links where
> necessary.
>   
>>> Using rdfs:seeAlso is as usual good but not precise enough. Maybe
>>> sameas.org could provide a minimal vocabulary to describe its URI,
>>> such as some http://sameas.org#hub property
>>> http://www.lingvoj.org/lang/fr     sameas:hub     http://sameas.org/
>>> html?uri=http://www.lingvoj.org/lang/fr
>>>       
>> Perhaps better:
>>
>> <http://www.lingvoj.org/lang/fr>
>>    rdfs:seeAlso
>>      <http://sameas.org/html?uri=http://www.lingvoj.org/lang/fr> .
>>     
> I think that cygri's email thread ended in a similar conclusion?
> I confess that the discussion has helped me to understand rdfs:seeAlso.
> I had tried to understand it through the semantics.
> But now I can see that from the point of view of the consumer it is just an
> optional
> #include <http://sameas.org/rdf?uri=http://www.lingvoj.org/lang/fr>
> (if you will pardon the C syntax!)
> That is, if I am following my nose bringing stuff into my local RDF cache to
> play with, and I hit one of these (?s rdfs:seeAlso ?o), I can decide to add
> the RDF resolved by ?o to the mix. In fact, I don't even care what the ?s
> is, other than as part of the decision whether to follow ?o or not.
> I guess everyone else already understood this. :-)
>   
>> <http://sameas.org/html?uri=http://www.lingvoj.org/lang/fr>
>>    rdf:type
>>      sameas:Hub .
>>     
> Not sure why this would be needed.
> But can be done.
> By the way, I think the URI should be:
> <http://sameas.org/?uri=http://www.lingvoj.org/lang/fr>
> To allow the conneg.
>   
>> As this will work in existing tools that understand rdfs:seeAlso.
>>     
> Which is always good.
> Best
> Hugh
>   


-- 

*Bernard Vatant
*Senior Consultant
Vocabulary & Data Engineering
Tel:       +33 (0) 971 488 459
Mail:     bernard.vatant@mondeca.com <mailto:bernard.vatant@mondeca.com>
----------------------------------------------------
*Mondeca**
*3, cité Nollez 75018 Paris France
Web:    www.mondeca.com <http://www.mondeca.com>
Blog:    Leçons de Choses <http://mondeca.wordpress.com/>
----------------------------------------------------**

Received on Tuesday, 9 June 2009 07:47:27 UTC