W3C home > Mailing lists > Public > semantic-web@w3.org > March 2005

Re: fast inferencing with jena and "?"

From: Danny Ayers <danny.ayers@gmail.com>
Date: Mon, 21 Mar 2005 10:31:01 +0100
Message-ID: <1f2ed5cd0503210131134dcbe6@mail.gmail.com>
To: Leo Sauermann <leo@gnowsis.com>
Cc: semantic-web@w3.org

On Sun, 20 Mar 2005 16:39:01 +0100, Leo Sauermann <leo@gnowsis.com> wrote:
> 
> We are using jena in a semi-commercial application and have to do simple
> RDF/S inferencing on subclasses and subproperties.

Any idea what algorithms are fast for this part of inferencing? I bet
there's something generic (non-RDF) that's fast and crude.

> Question:
> What inference engine can be plugged into Jena smoothely?

Offhand, I'm not sure, but there are lots of Prolog engines around for
Java, and one or two Rete engines. Actually I'd be surprised if
someone *hadn't* already got these plugged into Jena. Probably worth
trying the Jena mailing list ;-)

 I don't know if anyone's tried this yet, but it should be possible to
byte-compile Pychinko [1] for Java using Jython. Whether or not it
would be possible to operate directly on Jena models or whether you'd
need to serialize back and forth is another matter...

But if you find a good algorithm for the specific operations you need
you could always hard-wire it for performance.

Cheers,
Danny.

[1] http://www.mindswap.org/~katz/pychinko/

-- 

http://dannyayers.com
Received on Monday, 21 March 2005 09:58:04 UTC

This archive was generated by hypermail 2.3.1 : Tuesday, 1 March 2016 07:41:45 UTC