[CL] CL Technical Report

[John F. Sowa]

Leo,

We are planning to write a CL technical report,
which would discuss the motivation and use of
the various features of CL and their mapping
to and from various versions of logic.

Dan Gilman said that the TR could be considered
part of the CL project, and it would not require
a vote for a New Work Item.  Therefore, we can
begin to work on the TR now.  The result would be
published as an ISO TR, which would give it some
formal blessing.  However, most of us intend to
publish our chapters on the WWW as soon as they
are ready.  But the full final package would
become an ISO book.

I volunteered to be the editor of the TR, but
since I am not yet a member of X3L8, that is
just a statement of intention.  VivoMind intends
to become an X3L8 member, with me as the principal
representative to X3L8 and Arun as the alternate.

We still do not have a formal outline of the TR,
but following are some informal ideas.  If anybody
I mention has other ideas about what they would
like to do, please say so.  And if anybody thinks
that more should be added, please say so (with the
understanding that people who speak up might get asked
to volunteer unless they write a strong disclaimer).

Pat Hayes and Chris Menzel are planning to write
a chapter about the model theory and the motivation
for the various features.

Jay Halcomb and Randy Schulz are planning to write
a chapter about the relationships of KIF to CL and
guidelines for KIF users who want to translate their
KIF axioms to CLIF.

I plan to write a chapter on conceptual graphs, which
would go into more detail about how to use the various
features, their implications, and the differences
between a graph logic and a linear formula.  As part
of the historical discussion, I plan to mention Peirce's
existential graphs (which are roughly equivalent to the
untyped Core CGIF) and the various semantic networks of
the 1960s and 1970s.  See, for example, my chapter on
Semantic Networks from the Encyclopedia of AI (but I
don't plan to go into this much history in the TR):

    http://www.jfsowa.com/pubs/semnet.htm

As part of the CG chapter, I have to discuss the
differences between a typed and untyped logic, since
CGs are weakly typed (a type mismatch makes the statement
false instead of causing a syntax error).  I am planning
to use Z as an example of a strongly typed language.

We still haven't had a formal plan and outline for
the TR, so all of the above is still a very rough
guide, which will undoubtedly change as soon as we
finish the current CL draft and move into TR mode.

We should have more info about mapping CL to RDF(S),
OWL, and SWRL.  Pat worked on the model theory for
LBase with Guha, and that had a major influence on
the CL model theory.  More discussion of the SemWeb,
however, would be useful, but it is not clear whether
anybody will have time (or may volunteer) to do more.

I also don't think that I'll have the space to do a
full mapping from Z to CL in my chapter or just sketch
the mapping.  It would be good to have a separate
chapter on Z with more discussion of typing in general.
But that might not happen without another volunteer
to serve as the author or at least a coauthor.

This is just my summary of where we are and what we had
been talking about.  It is not yet a formal commitment,
but we should plan to make it more formal as soon as we
finish our current round on the standard.

It would also be good to have chapters on other topics,
such as the relationship of CL to logic programming and
related systems, including a discussion of the issues
about negation in Prolog, Datalog, and SQL.

And as soon as we talk about non-classical negation,
we have to discuss nonmonotonic logic, since negation as
failure is one of the most popular methods for supporting
nonmon logic.  And of course, the XML mapping is very
important for practical purposes.  And there are other
important topics, which could also be considered, such as
theorem proving, etc.  Maybe we could get somebody from
the UML community to write about the issues of defining
UML in Z or CL, etc.  That might get them more committed
to doing something.

But we need more volunteers, if any of the extras are
going to be done.

John