On Wed, 2011-02-23 at 15:20 -0500, Mike Spreitzer wrote: > So TDB+reasoner would also perform very badly, right?
Yes, though TDB is generally faster than SDB so it would help a little. Dave > > Thanks, > Mike Spreitzer > SMTP: [email protected], Lotus Notes: Mike Spreitzer/Watson/IBM > Office phone: +1-914-784-6424 (IBM T/L 863-) > AOL Instant Messaging: M1k3Sprtzr > > > > From: Chris Dollin <[email protected]> > To: [email protected] > Date: 02/23/2011 09:55 AM > Subject: Re: Reasoning in SDB > > > > On Wednesday, February 23, 2011 02:13:25 pm Mike Spreitzer wrote: > > Yes, I suspected there would be a way to create a suitable assembler > > description; my problem is that I have been unable to figure out what > that > > would look like. The assembler documentation I have been able to find > > does not discuss how to add a reasoner. What would such an assembler > > description look like? > > SDB assembler: > > http://openjena.org/wiki/SDB/Dataset_Description > > Assembler quickstart: > > http://www.openjena.org/assembler/index.html > > shows a sketch of an inference model spec, and the assembler manual > section on inference: > > http://www.openjena.org/assembler/assembler-howto.html#TOC-ID-7 > > (replace the database stuff in that with SDB models -- hmm, should > remove the old RDB stuff or at least delegate to an appendix) says > a little more. > > > I suspected the performance of SDB + OWL Reasoner would be very bad. > > Yes. > > > So my next question is: what's the easiest way to compute the OWL > closure of > > my model and then load into SDB? Can I do that by passing some suitable > > > assembler description to `sdbload` ? > > Load it into a memory model to do the inferences, then copy that model > into your SDB. This assumes that you have enough room in-memory to do > the inference. I don't think you can do it with just an assembler > description; > maybe some additions to content descriptions would do the trick ... > > Chris >
