[CL] (S)CL mapping for Object Role Modeling?
John F. Sowa
sowa at bestweb.net
Fri Nov 18 17:17:33 CST 2005
Thank you for telling us about this development.
My initial impression is very simple: it's going
to be a total, hopeless, unmitigated disaster if
it progresses in the direction it's going.
The most important point that I have to emphasize
about logic is its enormous power derived from
extreme simplicity. Peirce's existential graphs,
which I adopted as the basis for CGs, are based
on only three primitives: conjunction, negation,
and the existential quantifier. All of extended
CGIF is defined by mapping it to core CGIF which
uses only those three primitives. The Common Logic
core is somewhat larger than Peirce's, but still
not very much larger -- the whole semantics and
abstract syntax takes just 23 pages. That's all.
When I saw that the SBVR has over 400 pages, I was
very concerned, and when I looked at it in more
detail, my concern turned to sheer abject horror.
This is *not* the way to go.
I have no quarrel with defining terms like "policy"
and such, but they have to develop the standard
in stages. Please tell them to study the CL draft:
a simple abstract core and multiple languages on top.
Look at Annex B for CGIF -- a rich extended language
semantically defined by its mapping to core CGIF,
which maps to the CL core.
When I look at that document filled with an enormous
vocabulary of terms all at the same level, I know that
this is not going to work. They cannot do this project
without splitting it into a layered specification with
an extremely simple core. I would recommend Common Logic,
but if they would rather choose Z, I would be happy with
that as well -- because Z can very easily be mapped to CL.
I see that they want to do modal logic, but they are
going into areas where angels fear to tread. They are
not only supporting alethic modalities, they've also
got deontic modalities, and all kinds of policy crap.
Things like that can be done, but they're treading
on open research issues without the slightest clue
about what they're doing.
There may be some people who studied logic working
the authors, but I can't imagine anybody who knows
what he or she is doing who would condone such a
monstrosity. It looks like the committee took over
from anybody who knew what they were doing and dumped
a truck load of smelly fertilizer on a tiny plant.
I don't consider SBVR to be any kind of competitor
to Common Logic because it doesn't have the slightest
prayer of a hope of getting off the ground. But
that document shows that a lot of people in a lot
of corporations have poured a lot of time, energy,
and money down the toilet.
That is not the way to design a system of logic.
More information about the CL