Charles Coldwell <coldwell@[EMAIL PROTECTED]
> wrote:
> glen herrmannsfeldt <gah@[EMAIL PROTECTED]
> writes:
> > In a method related to the one Richard mentions, quad
> > could be written to accept a single variable, either a
> > C void pointer or unlimited polymorphic,
> > and pass this variable onto the called routine.
>
> That would be called a "thunk" or "continuation" in other programming
> languages ....
Thanks for the terminology pointer. I've heard of thunks, but hadn't
quite figured out what they were about. This gives me at least a hint
(and I'll assume that it might be an approximation to the definition, so
I'll avoid taking it to strictly). So I've been using thinks without
knowing it? Ok.
Reminds me of the epiphiany I had on listening to John Cuthbertson's
talk on a proposal for object oriented features in Fortran about a
decade ago. Prior to that, I had heard lots about object orientation,
and several of its basic concepts, but the examples just didn't "click"
with me. They struck me as contrived and unrelated to anything I did in
Fortran. But John's talk put it in a context that I could relate to.
Suddenly, I realized that I had long been programming using some of
those concepts. I just hadn't abstracted them sufficiently to recognize
them as distinct identifiable concepts. They were just part of the way
that I organized and modularized some of my code. "So that's what
inheritance is about!" It also hadn't occurred to me that a compiler
could actually help instead of getting in the way. And once I saw the
formalism, it helped me see ways to improve my coding in that area, even
before the features were aded to the language. Sort of like fairly
abstract mathematics can sometimes (indeed often) help guide you in the
solution of more concrete problems.
--
Richard Maine | Good judgement comes from experience;
email: last name at domain . net | experience comes from bad judgement.
domain: summertriangle | -- Mark Twain
|
