Fwd: [SWIPL] Just-in-time multi-argument indexing (take two)

[Chris Mungall]

I think this will be very good news for Thea - both for standard querying of the TBox, and also for code such as the rl reasoner.

Unfortunately I don't have any time to test using the latest swi in git, but will do next swi release

Begin forwarded message:

> From: Jan Wielemaker <J.Wiel...@vu.nl>
> Date: October 11, 2011 8:05:54 AM PDT
> To: SWI-Prolog maillinglist <swi-p...@lists.iai.uni-bonn.de>
> Subject: Re: [SWIPL] Just-in-time multi-argument indexing (take two)
>
> Hi,
>
> Small update. Some of you have played with the new just-in-time indexing
> code. Thanks for the reports. This also included two bug-reports. In the
> meanwhile I was making further changes to the indexing datastructures to
> prepare for indexing inside compound terms and wrote stress-testing
> code. This intermediate version fixes the reported problems. For this
> reason I decided to push it to the master.
>
> Most likely more internal changes will follow shortly because I'm in
> doubt whether some of the choices I made form a good basis for future
> enhancements to indexing. Be warned.
>
> In the meanwhile, I'm interested it programs doing lots of manipulation
> to dynamic predicates with many clauses that I can run as tests.
> Ideally, these programs run without output, produce a result that can be
> tested for correctness and have a runtime between seconds and -say- an
> hour. I have a few of those, but more is better.
>
> Cheers --- Jan
> _______________________________________________
> SWI-Prolog mailing list
> SWI-P...@lists.iai.uni-bonn.de
> https://lists.iai.uni-bonn.de/mailman/listinfo.cgi/swi-prolog