Re: Linked Data and IRI dereferencing (scale limits?)

Gunnar Aastrand Grimnes wrote:
> Joern's other problem here is that the types used by dbpedia include 
> things that are not covered by the dbpedia endpoint, i.e. opencyc, 
> skos, etc.

Where did you get the conclusion above from?

How is DBpedia distinct from its endpoint?

Have you looked at: http://dbepedia.org/fct ? It will give you a more 
holistic view of this data space. There are linkbases in place via 
distinct named graphs.


>
> Now he has to discover the sparql endpoint for these.
>
> Luckily VOID + rkbexplorer has the solution for him:
>
> http://void.rkbexplorer.com/endpoint-search/
>
> Keith points out a more elegant (i.e. no specific webservice) 
> solution, but it doesn't work everywhere: 
> http://blogs.talis.com/n2/archives/914
Really don't grok you conclusions here at all.


Kingsley
>
> Cheers,
>
> - Gunnar
>
> On 06/08/10 21:51, Giovanni Tummarello wrote:
>> Only solution for you now is to use SPARQL instead of resolving the URI.
>>
>>         Much less traffic and it would actually work
>>
>>
>>     SPARQL doesn't make the problem go away, it just pushes the limits
>>     further out. SPARQL endpoints that see significant traffic have
>>     similar restrictions built in, either on query complexity or query
>>     runtime or number of results. So you might hit the limit at 16000
>>     statements rather than 2000 or whatever.
>>
>>
>>
>> Jorn could have asked for the rpoperties he knows how to handle and
>> would have received them (his problem was that they were cut off by the
>> 2000 limit)..
>>
>> Gio
>
>


-- 

Regards,

Kingsley Idehen	      
President & CEO 
OpenLink Software     
Web: http://www.openlinksw.com
Weblog: http://www.openlinksw.com/blog/~kidehen
Twitter/Identi.ca: kidehen 

Received on Wednesday, 11 August 2010 13:12:22 UTC