Readers should note that LUBM, as generally used, is not expressible in DL-Lite because it uses transitive properties and because qualified existentials are used in the left side of subsumption relations. The results presented here are with a modified version that reduces the expressivity.
JohnGoodwin
· 1 year ago
Is there a demo version we can download and experiment with?
Kendall
· 1 year ago
John,
There will be a first public release soonish; we'll be giving demos at OWLED next week, and I think some Ordnance Survey folks will be there. After that, we'll make a public release.
lolive
· 1 year ago
IBM is working on a DLP (and now DL) engine on top of DB2. there is an attempt from the EPFL to add such support to Oracle. OwlGres is targetted at Postgres.
Definitely there is something to investigate, it seems :)
From a developper point of view, I have a concern. Is it possible to abstract such a technology from the RDBMS? So we are not tied to a given "vendor".
Or does it require fine tuning of the database internals?
Kendall
· 1 year ago
Olivier,
I can't speak for the other projects, but Owlgres isn't Postgres-specific or even tied to Postgres. We developed it with Postgres in mind, but we don't any Postgres internals tweaking. Owlgres exists purely at the JDBC level, except for database-specific setup code.
So we could easily make Owlgres work for Oracle or DB2, and probably will do so at some point.
To get the *most* performance, one might tweak some DB internals (which is one reason we *prefer* Postgres), but so far that's not been necessary.
Oleg Bartunov
· 1 year ago
Markus,
I'm PostgreSQL developer and would like to work on effective RDF storage in PostgreSQL in the framework of our astronomical project "Astronet" (www.astronet.ru). We developed and maintain GiST and GIN indexing framework as well as text search and other PostgreSQL extensions. I'm not an expert in ontology technologies yet and have no practical experience. If you have any specific postgreSQL problems or ideas how to improve performance of semantic queries, don't hesitate to communicate.
All Comments
There will be a first public release soonish; we'll be giving demos at OWLED next week, and I think some Ordnance Survey folks will be there. After that, we'll make a public release.
there is an attempt from the EPFL to add such support to Oracle. OwlGres is targetted at Postgres.
Definitely there is something to investigate, it seems :)
From a developper point of view, I have a concern.
Is it possible to abstract such a technology from the RDBMS?
So we are not tied to a given "vendor".
Or does it require fine tuning of the database internals?
I can't speak for the other projects, but Owlgres isn't Postgres-specific or even tied to Postgres. We developed it with Postgres in mind, but we don't any Postgres internals tweaking. Owlgres exists purely at the JDBC level, except for database-specific setup code.
So we could easily make Owlgres work for Oracle or DB2, and probably will do so at some point.
To get the *most* performance, one might tweak some DB internals (which is one reason we *prefer* Postgres), but so far that's not been necessary.
I'm PostgreSQL developer and would like to work on effective
RDF storage in PostgreSQL in the framework of our astronomical
project "Astronet" (www.astronet.ru). We developed and maintain
GiST and GIN indexing framework as well as text search and other
PostgreSQL extensions. I'm not an expert in ontology technologies
yet and have no practical experience. If you have any specific postgreSQL problems or ideas how to improve performance of semantic queries, don't hesitate to communicate.
Best regards,
Oleg