Re: Documentation of PowerLoom reasoning mechanism

Sebastian,

I am sorry to say that we've never theoretically analyzed PowerLoom (since =
we
are always too busy extending it :-)  So there is no hard information on th=
at
anywhere.  Conceivably, we could do that for the strict first-order subset =
of
the language (that is without all the hairy stuff such as selective closed-=
world
reasoning, defaults, holds, etc.), but then it's hard for us to justify
spending resources on that.

A natural deduction inference rule PowerLoom doesn't implement is reasoning=
 by
cases, that is:

   A |- C
   B |- C
   A or B
   ------
     C

The reason is that there doesn't seem to be a good way to focus this kind
of reasoning in a natural deduction system except for certain special cases.
I am not aware of any other inference rule we are not implementing.

There are also currently some limitations in normalization such as
skolemization and with complex nested quantification.  Some of these have b=
een
resolved in our local development copy, some of them have been pending for a
long time due to the lack of a strong motivator to fix them.

While these limitations are certainly inelegant, they can usually be worked
around.

Hope that helps,

Hans

> From: Sebastian Herold <[email protected]>
> Sender: [email protected]
> To: [email protected]
> Subject: [PowerLoom Forum] Documentation of PowerLoom reasoning mechanism
> Date: Mon, 16 May 2011 16:18:04 +0200
> MIME-Version: 1.0
> Content-Type: text/plain; charset=3D"iso-8859-15"

> Hi there,

> I'm currentling looking for a more detailed description of PowerLoom's =

> reasoning functionality regarding its theoretical incompleteness. =

> Somewhere in the documentation it is stated that not all rules of =

> deduction are supported. Can I find somewhere information about

> 1) which rules are (not) supported and/or
> 2) examples of queries that could hence not be answered by PowerLoom?

> Thanks in advance,
> Sebastian
> -- =

> Dipl.-Inform. Sebastian Herold
> Technische Universit=E4t Clausthal
> Institut f=FCr Informatik - Software Systems Engineering
> Lehrstuhl von Prof. Dr. Andreas Rausch
> Julius-Albert-Str. 4
> 38678 Clausthal-Zellerfeld
> Telefon: +49 5323 72-7164 - Sekretariat: -7160
> Telefax: +49 5323 72-7169
> Raum: Albrecht-von-Groddeck-Str. 7, Raum 202
> WWW: http://sse.in.tu-clausthal.de