[Attempto] Should I allow for non-monotonic reasoning?

[Douglas Miles]

Great thank you!  Please do it sooner rather than later.  I am working on a
similar system in (SWI) Prolog ..  And will be happy in sharing some
overlap.  So we don't end up having to just replicate each others work.. .
I am doing negation as explicit proving.. Using much as modus tollens logic
in unit projection (negation always yields bindings).  This has made
negation by failure a completely unneeded.  But still in ways can leverage
it

some notes:

 During inference there are times I can get neither the positive or
negative proof of a literal (what i call unknown) which can be presented to
the user later that they may find that is an appropriate area for knowledge
collection.

 Every back-chain requires I do a complete negative and positive literal
query in the antecedents.

 To implement this I use a combination of ... Mark Stickel's PTTP (
https://github.com/TeamSPoon/PrologMUD/blob/master/pack/logicmoo_base/prolog/logicmoo/logicmoo_engine.pl)
method to get  modus tollens not_p horn clauses by converting the logic
(once CNF) to in horn clauses however those I leave subject to a
memoization in SWI-prolog created by Feliks Kluzniak (
https://github.com/TeamSPoon/PrologMUD/blob/master/pack/logicmoo_base/prolog/logicmoo/dra/dra_common.pl
)

 I use Tim Finin's PFC (prolog forward chaining
http://www.csee.umbc.edu/~finin/pfc/src/  ) library on the positive
yielding Modus Ponens (and produce skollems) (when it isn't too expensive)

 If all this makes sense to you .. is this similar .. do we have overlap?




On Wed, Aug 12, 2015 at 1:52 AM, Norbert E. Fuchs <fuchs at ifi.uzh.ch> wrote:

>
> > On 12 Aug 2015, at 9:58 , Douglas Miles <logicmoo at gmail.com> wrote:
> >
> > What language is RACE in?
> >
> > /me hopes its prolog!
>
> Douglas
>
> You are hoping correctly. RACE is completely written in (SWI) Prolog.
>
> Best.
>
>    --- nef
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.ifi.uzh.ch/pipermail/attempto/attachments/20150812/58fafef9/attachment.html>