Re: [tlug] Functional Programming Group Meeting

["SL Baur" <steve@example.com>]

On 4/19/08, Curt Sampson <cjs@example.com> wrote:
> On 2008-04-18 00:03 -0700 (Fri), SL Baur wrote:
>
>  > What do you think of Icon?
>
>
> I had a look at it the other day, actually, having been led there by a
>  few moments of SNOBOL nostalgia.

For those of you just tuning in, Icon was done by the same person who
did SNOBOL and is properly considered the successor to SNOBOL.

>  I like it for what it is, but I have
>  to say that since about the mid 80s Algol-style syntax has seemed a
>   bit "heavy" to me, and that put me off it a bit. It's kind of like the thing
>  where you do C for a few years and go back and look at PL/1 again, and
>  it's just not quite as cool as it used to be.

That's probably the major failing of Icon.  It's very easy to write code
that looks like C.  It isn't.  Big time.

For getting stuff done fast, Icon is great.  I was experimenting with it
long before I had an opportunity to use it in a classroom setting.  The
first homework assignment I had the chance to use Icon in in a
Functional Programming class I had, I had the right answer in a matter
of minutes.  It took me several hours of debugging and tracing to prove
to myself that it really was the right answer because the professor
had given us an impossible problem to solve and the point of the
assignment was to prove that it was impossible.  I had other similar
experiences later.

>  The one syntax (or lack therof :-)) that has always and probably will
>  always work for me is LISP. It's amazing to look back and realize that,
>  as it were, "everything I needed to know I learned in the 50s."

Lisp is *the* most successful language ever designed.  The syntax
has withstood the test of time.  It's trivial to learn.  In another half
decade (the same amount of time Lisp has been around), Java, C++
will be regarded as fads of the late 20th and early 21st century.
Meanwhile, folks will still be writing Lisp code for emacs.

>  (Ok, I'm
>  romanticising; it wasn't until Scheme came out that we actually merged
>  the namespaces for variables and functions.)

Scheme and Lisp are as equivalent as the FBAPP[1] languages (Fortran,
BASIC, Algol, Pascal, PL/I).

[1] That's what my advisor and compiler design professor at Caltech
called them.  C was "sugar coated assembly" and also FBAPP.