W3C home > Mailing lists > Public > public-owl-dev@w3.org > January to March 2011

Re: ANN: SPARQL-DL query engine for the OWL API

From: Holger Knublauch <holger@topquadrant.com>
Date: Sun, 6 Mar 2011 09:01:05 +1000
To: public-owl-dev@w3.org
Message-Id: <70353801-08A5-45E7-A8AE-371E07CF1A71@topquadrant.com>

On Mar 6, 2011, at 7:20 AM, Thorsten Liebig wrote:

> Michael Schneider wrote:
>> Holger Knublauch wrote:
>>> Hi Thorsten (and others),
>>> have you considered that SPARQL-DL may not be a good name for a
>>> technology that has an entirely different syntax and semantics from
>>> SPARQL? When I looked at this, I expected to find a DL-friendly subset
>>> of SPARQL. But this is not what SPARQL-DL is.
>>> I believe there is enough confusion on the marketplace already. Why not
>>> call it something like OWL-QL?
>> Because there already is an "OWL QL" query language:
>>     <http://www.ksl.stanford.edu/projects/owl-ql/>
>> ... and because there is now also an "OWL 2 QL" ontology language:
>>     <http://www.w3.org/TR/2009/REC-owl2-profiles-20091027/#OWL_2_QL>
>> Good names are rare and precious. ;-)
> Absolutely right.

My suggestion was "something like OWL QL" - I do know that this particular name was already taken, but this is just a red herring. It should be very doable to find a more appropriate name. Is DLQL already taken?

> Furthermore, the term "SPARQL-DL" is not our invention. The name has its seeds in a OWLED '07 paper from Sirin and Parsia that introduced a subset of SPARQL for OWL DL. Our implementation basically is an extension of this query language wrt. OWL 2 on top of the OWL API.

If the original SPARQL-DL was still a subset of SPARQL then this used to be an appropriate name. However, none of the example queries on your SPARQL-DL page are valid SPARQL, e.g.

     WHERE { Transitive(?p),
             PropertyValue(<http://example.com#myClass>, ?p, ?x) }
  PREFIX wine: <http://www.w3.org/TR/2003/PR-owl-guide-20031209/wine#>
   SELECT ?i
     WHERE { Type(?i, wine:PinotBlanc) }
     OR WHERE { Type(?i, wine:DryRedWine) }
This is no longer a subset of SPARQL, and calling it SPARQL-DL will IMHO cause problems and be counter-productive for all parties involved. As a tool vendor we will be facing questions as to whether our SPARQL support is incomplete because we don't support "SPARQL-DL".  And you will be asked about support for aggregations, sub-selects, built-in functions and all other features of SPARQL 1.1. Do you really want to be answering all those questions for the rest of your life? Only because you use SELECT and WHERE does not make it SPARQL.

Thanks for your consideration
Received on Saturday, 5 March 2011 23:01:47 UTC

This archive was generated by hypermail 2.3.1 : Tuesday, 6 January 2015 20:58:20 UTC