[SCL] Representing constraints that go beyond EBNF

[John F. Sowa]

Murray,

I think we're converging.  There are fewer points to discuss.

 > [EBNF] ...  Would that specific subset have a name? A particular
 > domain of usability? Or would it simply be a subset because
 > it was the part of SCL that was defineable via EBNF?

It can just be called the "EBNF subset".  The subset of logic that
supports it is called Horn-clause logic, which is the subset used
in logic programming languages, such as Prolog.  Besides pure
logic, it would require an ontology for sequences, since every
language defined by EBNF consists of sequences of characters or
tokens, and the result of the parse is a tree structure, which
can also be expressed as a sequence of sequences (or list of lists).

To define the sequences, I would recommend that we use the axioms
for sequences that are already defined in the ISO standard for Z.
Since it's a standard, we can just refer to it.  The Prolog
community has already specified the basic approach in their
definite-clause grammars (DCGs).  The only work needed is to
pick up those definitions and adapt them to the SCL notation.

 > ... If at some point KIF becomes an ANSI or ISO standard
 > we could reconsider and write an Annex.

The SCL standard will also standardize the concrete syntaxes of KIF,
CGIF, and infix predicate calculus.  So they would all be standards.
Infix predicate calculus would probably be familiar to more readers
than the others.

 > If there's some agenda to bringing the SCL and Z work together,
 > that should probably be taken up as a separate standards activity
 > under the same aegis as this work, but after we're finished.

Z is already and ISO standard and the question has been raised about
how it is related to SCL.  Defining Z in terms of the SCL abstract
syntax is not difficult, and it would be a useful demonstration.
The only questions are how big we want to make the standard, how
much time we have to write it, and who will do the writing.

 > I think you also have to realize that (a) people use different
 > terminology than you do if they come from a different domain, and
 > (b) people's expertise is often not in what you think it is, even
 > based upon their ostensible qualifications and background. Mine is
 > a case in point...

No quarrels here.  I am very happy to help people who want to broaden
their horizon by branching out into a new field.  But I have been
frustrated with so many people who go into a new field without bothering
to spend some time learning the basics of that field.  Even worse,
many of those people become Pointy-Haired Bosses, project leaders,
project, paper, book, or grant reviewers, and pundits who make
pronouncements about things they don't understand.

So the tone of voice that underlies most of my complaints is a
groaning "Oh, no!  Here we go again......"  If I step on any toes
that don't deserve to get stepped on, just squawk, and I'll apologize.

John