Message 12714 of the SUO list

Subject: Re: Time to request PAR extension
Date: Tue, 24 Aug 2004 01:07:23 +1000
From: Philippe Martin
In-reply-to: msg12712 by John Sowa


> For the MSO, the most important question is
> how close to completion is the specification
> of the framework and the methodology for
> aligning, merging, and maintaining ontologies.

I interpret that as a question about the support provided
by WebKB-2 (or its future open source version, AnyKB).

I do not plan to extend much the protocols which regulate
the update of the KB by multiple users (and which force a
user who introduces a new statement/belief to relate it to
the beliefs which it specializes, generalizes or
contradicts).
I plan to extend the syntax of FCG and FT to propose a
more intuitive and "normalizing" format for certain features
from KIF (then their handling will have to be implemented
too).
The import and export procedures from/to KIF+OWL, KIF+SUMO,
RDF+OWL and others are at their beginnings, so I have quite a
lot of work to do there. This includes the removal of
constraints on knowledge entering (e.g. the declaration of
categories before their uses in statements,  the use of
strict subtypeOf links rather than subtypeOfOrEqual links,
the absence of redundancies, etc.) which are enforced in WebKB-2
(for knowledge checking and normalization purposes) but not
respected in RDF/KIF ontologies.
So far, I had not planned for WebKB-2 to support any automatic
"manipulation (comparison, ...) of ontologies as objects"
as is in IFF for example (for details and rationales, see
http://www.webkb.org/doc/onSUOlist/m040531phmartin.html)
but I'll gain more perspectives on this issue at the
LOA.
There are also quite a few bugs to iron out and procedures to
implement more "fully".

So enough work for a year (especially since I won't work on those
particular points full time) but I do hope that AnyKB will be
released, and usable by people other than me, sooner than that.

Philippe