W3C home > Mailing lists > Public > public-lod@w3.org > June 2008

Re: The king is dressed in void

From: Andreas Langegger <al@jku.at>
Date: Tue, 17 Jun 2008 22:47:40 +0200
To: Giovanni Tummarello <giovanni.tummarello@deri.org>
Message-Id: <061DCC65-31A7-4BEA-AD4B-A374DCBE380F@jku.at>
Cc: Michael Hausenblas <michael.hausenblas@joanneum.at>, public-lod@w3.org, Semantic Web <semantic-web@w3.org>


On Jun 13, 2008, at 1:27 AM, Giovanni Tummarello wrote:
> Licence: yes i agree, it will be added to the sitemap extention much
> like it happens in microformats already.
> if you want to use RDF i believe this is what you're looking for
> http://validator.creativecommons.org/

yes, especially for LOD. In our case, we'll have to provide means of  
restriction and even billing (when SemWIQ gets integrated into Grid  
middleware - our requirement is a middleware for virtual data  
integration in the Grid). However, because I want SemWIQ being as  
generic as possible and also usable as a general data integration  
middleware, I want to re-use as much as possible and build on  
standards to be compatible to LOD/WOD.

> Statistics:
> i'd tend to see this use case as a low level one that concerns
> implementation of distributed sparql (an interesting aspect however!).

sure. That's why I would extend voiD and maybe even Scuvo [1] for my  

> It seems strange at least to ask people to write some triples to say
> how many triples they have when a sparql endpoint is there just to
> answer you any query you might want?

that's why I said, it's just because of performance.
At the moment my DataSourceMonitor uses a { [] a rdfs:Class } query to  
get classes of a dataset and then it iterates over classes to fetch  
instance statistics: count (*) where { [] a c_i } - this can be done  
remotely. However, for good cost-based optimization I need histograms  
for any class/property[/range] combination. RDFStats will run directly  
(and "nice") at a SPARQL endpoint to generate such histograms - at  
least in my case. Query federation for >20 public SPARQL endpoints is  
not feasible anyway with acceptable performance if you have many joins.

Sorry for the late answer,

[1] http://sw.joanneum.at/scovo/schema.html

Dipl.-Ing.(FH) Andreas Langegger
Institute for Applied Knowledge Processing
Johannes Kepler University Linz
A-4040 Linz, Altenberger Straße 69
Received on Tuesday, 17 June 2008 20:48:32 UTC

This archive was generated by hypermail 2.3.1 : Wednesday, 7 January 2015 15:15:50 UTC