[SCL] SCL vs. Z(ed)

[John F. Sowa]

Pat and Harry,

I don't believe that either of you need to worry about Z.
On the contrary, I believe that the ISO Z standard will help
support the CL effort in several ways.  Among other things,
the Z workbench with axioms for sets, bags, sequences, and
integers would be a very useful supplement to CL.  For the
CL standard, we can just refer to the Z workbench as an
optional supplement.

But Z is less expressive than CL and it cannot serve the
purposes for which CL has been designed.  First of all, Z is
a strongly typed language, which cannot be used as a language
into which nontyped languages, such as RDF, full OWL, untyped
KIF, and ordinary predicate calculus can be translated.  This
point is addressed in the slides that Chris, Pat, and I prepared 
for the ISO meeting in Santa Fe:

   http://www.jfsowa.com/talks/clprop.htm

However, it is possible for an untyped language such as CL
to serve as the semantic foundation for a typed language
such as Z.  All of these issues have been addressed, see
the above slides for a summary.

Second, Z was designed to serve a different purpose:  to be
a *specification* language rather than a language to support
translations to and from a wide range of other languages.
CL can support translations to and from Z and other languages,
but Z cannot support translations to and from CL since Z is
not as expressive as CL.

Note to Harry:  Please forward this note to the L8 committee
to inform them that the SCL group is very well aware of Z and
will be supporting Z as well as other logic-based languages
that Z cannot support.  And for further info they should
consult the slides above.

John