Search Mailing List Archives
[protege-discussion] [Pellet-users] Big performance difference with SWRL when resolvingOWL imports
Tania Tudorache
tudorache at stanford.edu
Tue Dec 11 18:10:17 PST 2007
Serge,
This is the wrong list to post a Protege-OWL related question. Please
use in future the protege-owl mailing list.
I also wonder what Protege version did you use to do the tests? Protege
3 or Protege 4?
The latest beta of Protege 3.4 does not use the SWRL rules for the
reasoning. We hope to have this available very soon.
The reason why SWRLTab breaks with dl-safe.owl is that the file does not
follow the SWRL specification when building the atom lists: it uses rdf
lists instead of atom lists. If you have more questions about this,
please post them on the Protege-OWL mailing list.
Cheers,
Tania
Serge Libotte wrote:
> Funny, I just tried this.
> The perf is globally a bit better:
>
> StopWatch '': running time (millis) = 1562
> -----------------------------------------
> ms % Task name
> -----------------------------------------
> 01078 069% Owl Loading
> 00000 000% get order1
> 00000 000% get hasPart
> 00297 019% get needs
> 00187 012% get NeededRes
>
> On the top of that, Protégé doesn't complain about anything (which
> sounds normal to me since any specific XML editor is supposed to know
> the NS it should manipulate)
>
> I wonder why Protégé breaks in opening the SWRL tab with
> http://owldl.com/ontologies/dl-safe.owl
>
> 2007/12/7, raphael.coulonvaux at belgacom.be
> <mailto:raphael.coulonvaux at belgacom.be>
> <raphael.coulonvaux at belgacom.be <mailto:raphael.coulonvaux at belgacom.be>>:
>
> Suppress the include.
>
>
> ------------------------------------------------------------------------
> *From:* pellet-users-bounces at lists.owldl.com
> <mailto:pellet-users-bounces at lists.owldl.com>
> [mailto:pellet-users-bounces at lists.owldl.com
> <mailto:pellet-users-bounces at lists.owldl.com>] *On Behalf Of
> *Serge Libotte
> *Sent:* 07 December 2007 10:29
> *To:* pellet-users at lists.owldl.com
> <mailto:pellet-users at lists.owldl.com>
> *Subject:* [Pellet-users] Big performance difference with SWRL
> when resolvingOWL imports
>
> Dears,
>
> I've made a very small ontology using Protégé. (see attached)
> In it, I've defined two rules.
> When doing that, Protégé wants to import some other namespaces
> definitions, as you know.
>
> Then I wrote a simple java code to test expected results and they
> are as expected. Nice!
>
> owl loaded...
> Order_1 has hasPart(s):Spec_1, CFS_1, Offer_1
> Order_1 has needs(s):Res_1
> NeededRes instances: Res_1
>
>
> But...
>
> The application throws a lot of messages telling it cannot parse
> the imported owl (normal since I prevent it to access the internet)
> I then copied the imported owl locally and added some
> SimpleURIMapper to resolve NS to those one.
>
> The results are the same but the performance is bad: it takes 10
> times longer to retreive result of the rules (see below)
>
> StopWatch 'no mappers': running time (millis) = 1890
> -----------------------------------------
> ms % Task name
> -----------------------------------------
> 01359 072% Owl Loading
> 00000 000% get order1
> 00000 000% get hasPart
> 00328 017% get needs
> 00203 011% get NeededRes
>
> StopWatch 'with mappers': running time (millis) = 16812
> -----------------------------------------
> ms % Task name
> -----------------------------------------
> 01156 007% Owl Loading
> 00000 000% get order1
> 00000 000% get hasPart
> 08422 050% get needs
> 07234 043% get NeededRes
>
> Since results are the same I wonder:
> 1) is it necessary to import those NS?
> 2) how to work in Protégé without all those imports?
> 3) How to keep imports and avoid error messages while keeping good
> performances?
> 4) What makes the rules going so slow (ontology size?)
>
> Thanks for helping,
>
> Serge.
>
>
> _______________________________________________
> Pellet-users mailing list
> Pellet-users at lists.owldl.com <mailto:Pellet-users at lists.owldl.com>
> http://lists.owldl.com/mailman/listinfo/pellet-users
> _______________________________________________
>
> Sponsored by Clark & Parsia, LLC http://clarkparsia.com/
>
>
> ------------------------------------------------------------------------
>
> _______________________________________________
> protege-discussion mailing list
> protege-discussion at lists.stanford.edu
> https://mailman.stanford.edu/mailman/listinfo/protege-discussion
>
> Instructions for unsubscribing: http://protege.stanford.edu/doc/faq.html#01a.03
>
More information about the protege-discussion
mailing list