RE: XSB prolog's SQL/ODBC interface -- anyone have experience of it?

Dan,

You may wish to check out Virtuoso http://www.openlinksw.com/virtuoso

Virtuoso is an SQL92 database that allows creation of real-time XML views
against multiple back end relational database tables concurrently.  E.g. you
can evaluate a heterogenous join against SQLServer2000 on Win32 and Oracle
8i on linux, and return the whole thing as a dynamic XML doc.

Also, there is a built-in XSLT engine, so you can output the XML in any
format you wish.

Please feel free to follow up with any questions.

Best regards,
Andrew
----------------------------------------------------
Andrew Hill
Director Technology Evangelism
OpenLink Software
http://www.openlinksw.com
XML & E-Business Infrastructure Technology Provider

> -----Original Message-----
> From: www-rdf-interest-request@w3.org
> [mailto:www-rdf-interest-request@w3.org]On Behalf Of Dan Brickley
> Sent: Monday, January 29, 2001 4:29 PM
> To: www-rdf-interest@w3.org
> Subject: XSB prolog's SQL/ODBC interface -- anyone have experience of
> it?
>
>
>
> Hi all
>
> I'm hoping someone here might be able to help me out a little. I'm looking
> at XSB, a logic programming and deductive database system (see
> http://xsb.sourceforge.net/ ), and trying to figure out feasible it might
> make use of some of its capabilities in an RDF/SW context.
>
> Specifically, XSB can "out-source" data storage to good old fashioned
> relational database tables. Detailed at
> http://www.cs.sunysb.edu/~sbprolog/manual2/node90.html and my particular
> interest lies in the 'view level' interface documented at:
> http://www.cs.sunysb.edu/~sbprolog/manual2/node90.html
>
> The basic "relation level" interface in XSB allows you tell tell it that
> information about certain predicates is held remotely in an SQL-based
> database. The more sophisticated "view level" interface provides for more
> efficient use of the SQL store, by rewriting complex expressions into
> relatively efficient SQL, rather than slurping much of the data out of the
> RDBMS into Prolog before querying it locally.
>
> [[
> The view level interface can be used for the definition of rules whose
> bodies includes only imported
> database predicates (by using the relation level interface) described
> above and aggregate predicates
> (defined below). When they are invoked, rules are translated into complex
> database queries, which
> are then executed taking advantage of the query processing ability of the
> DBMS's.
> ]]
>
>
> My question is: this sounds great, how well does it work in practice? One
> could imagine using a very similar technique to wrap RDBMS data in an
> RDF/SW layer. I've only made the most basic of experiments with XSB (on
> Win32 using MS Access), so would love to hear from anyone on RDF IG who
> has made a more detailed study. I understand similar facilities are
> available for SWI-Prolog, and the technique of course is an idea that
> crops up in other contexts. Basically I'm looking for readily deployable
> techniques and tools that might  (for example) be used to add RDF smarts
> to Intranet applications, and make use of rather than replace existing
> data storage and management strategies. The ODBC facilities in XSB have
> long intrigued me, so I figured www-rdf-interest might be a good place to
> find out if this is a good lead...
>
> thanks for any pointers,
>
> Dan
>
> ps. if any listmember is in a position to host student projects (MSc etc)
> in this or related areas (eg. RDF visualisation) drop me a note...
>
> --
> mailto:danbri@w3.org
>
>

Received on Monday, 29 January 2001 16:56:30 UTC