William Tanksley
Fri, 16 Oct 1998 16:50:45 -0700

Francois-Rene Rideau wrote:

> > Thinking in lambda calculus is not natural.
> Is thinking natural at all?

Well put.

> I don't think I'm "sold" to lambda-calculus (as I've read in another message
> in this list). Perhaps there are simpler concepts that subsume
> lambda-calculus and give better insight into programming languages;
> if there are, I'll be glad to learn about them. I haven't seen them yet,
> and lambda-calculus is so simple that I'd be surprised if/when I meet them.

New things are always suprising.  Almost always; sometimes their unsuprisingness is
the thing that suprises me.

Take a look at Abstract State Machines (Evolving Algebras) for an example right in
your department (proof of correctness, translation from concept to code).  I for
one consider them to be more useful than the lambda calculus, in that they adjust
to how the machine works as well as how the algorithm works (and that adjustment
can be tuned by the evaluator).