W3C home > Mailing lists > Public > semantic-web@w3.org > April 2009

Re: [ontolog-forum] [STI-community] FINAL CALL: charting the?futureof semantics - STI 3rd Roadmapping Workshop

From: John F. Sowa <sowa@bestweb.net>
Date: Mon, 06 Apr 2009 18:58:10 -0400
Message-ID: <49DA8902.2090602@bestweb.net>
To: "[ontolog-forum]" <ontolog-forum@ontolog.cim3.net>
CC: Semantic Web <semantic-web@w3.org>
Dick,

Just a correction about what we object to:

RHM> John Sowa, Pat Hayes, Chris Menzel, et al,
 > have "objected" to using mKR because of
 > (1) no formal semantics theory
 > (2) no advantage over other Controlled English languages
 >
 > I readily admit (1).

Re #1:  There is no shortage of notations that have a formal
semantics, and it is very easy to specify a formal semantics
for mKR (at least the sentence part) by defining a mapping to
some notation that has a formal semantics defined for it
(and then asserting that the semantics of mKR is defined
by that mapping).

RHM> I don't agree with (2).  I contend that mKR's use of
 >   context and hierarchies
 >   genus-differentia definitions
 >   actions - commands - n-ary relations
 >   UNIX-shell-like variables and control structures
 >   make mKR more attractive and very practical.

What people prefer for aesthetic reasons is a personal
matter, so I would encourage anybody to choose whatever
notation they prefer.  In fact, that is one of the major
reasons why Common Logic has *no* preferred notation.
The ISO standard presents three distinctly different
dialects, and it encourages anybody to map their favorite
syntax to the CL semantics.

As I said before, the option of mapping to Unix commands
is very useful in any tool, and it's one aspect of the
mKE system that could make it quite practical.

So far, I have not seen a single item of the mKR declarative
parts (sentence and what you call context) that cannot be
mapped to CL or the IKL extensions.  But we have consistently
complained about the vagueness of what you call "context".

I am sure that if you ever pinned down exactly what you
mean, it would be easy to map the context features to IKL.

But instead of defining what you mean, you just pile up more
and more clouds of vague verbiage.  We have been trying to
point out where the vagueness lies, but you just pile up
the verbiage higher and deeper -- but no clearer.

John
Received on Monday, 6 April 2009 22:58:49 UTC

This archive was generated by hypermail 2.4.0 : Tuesday, 5 July 2022 08:45:11 UTC