Re: Required common programming language

[Leigh Palmer <palmer@sfu.ca>]

I'm being a Devil's advocate here, I know, but in this case I'm really
on the Devil's side. Let me comment on just three of Frank's parries:

> Leigh writes (in part, I'll edit to take him out of context)
>
> > ...and that some attention
> > should be paid to the virtue of the inclusion of the GO TO statement in
> > any reasonable language. Furthermore they should be told that no one is
> > ever going to read their code,
>
> There is no virtue to the GOTO.  I wrote a technical word processor in Atari
> Basic with machine language routines in critical spaces for speed.  I
> couldn't
> read my own code today to save my ... life!  It was spaghetti.  It needed
> to be
> compact because of space and speed considerations that are no longer
> necessary.
> Structured programming with clean modules provides reusable modules and the
> ability to re-use an old wheel from time to time.  If you can't read it 3
> months later, you can't modify it for a new task AND it may well have done
> some
> things you couldn't see buried in the spaghetti.

I omitted an important word there, and I apologize if it obscured my
meaning. I meant, and should have said "some attention should be paid
to the virtue of the inclusion of the GO TO statement in any reasonable
*first* language." That was what I thought we were discussing, and that
is the thrust of my piece.

It might happen that one writing a first computer program has no need
to reuse that code. Might I suggest that even you had no need to reuse
the code you wrote in machine language for the Atari? [Brag warning]
The same son I told you about has been using Macintoshes for 14.5* years
and is very talented. While working for Bill Gates he once made an
improvement in an early version of Word that chopped 75% off the refresh
time in one of them. As I said, he learned by an utterly different route
which was similar to my own, and he now writes beautiful structured code
in C++ and Java, only as a diversion from his real work, I'm pleased to
be able to say.

> > We old codgers never took programming courses, and still we
> > learned. I got a copy of McCracken and a key punch and went to work. It
> > didn't take long to get that first thrill, I'll tell you. Why should
> > others be deprived of it by having unnecessary obstacles placed before
> > them because they are so unfortunate as to have to learn structured
> > programming and to spend time writing grammatically correct documentation
> > for the eyes of a teacher?
>
> I learned it the same way.  That doesn't mean it was the best way, just
> what we
> did in those earlier times.  Having lived through that time AND enjoyed
> programming as you did, I still enjoy it, but the structure gives it an
> additional beauty with the practical result that I can even read it a few
> months later.

I like to think that my construction of English prose is superior to my
writing in elementary school. I probably write things today that I would
have been unable to read when I was eight years old (or at least I hope
I do). That doesn't mean that I would have got to this stage of writing
faster, or written better, had I learned to write correctly when I was
five years old.

I am not saying that structured programming is in any way "bad." I wrote
structured programs in APL and Fortran II; I know no language in which
I can't write structured programs, and certainly structure has all the
advantages you claim for it. I'm saying that I don't think it is
necessary to start out programming that way, and that the GO TO statement
(the elimination of which was the proudest achievement of Pascal) is a
perfectly good, useful tool. So is the word "ain't"!

> > The point here is this: computer programming is not necessarily something
> > that has to be taught. Like other language, with a facility for
> > communication (access to other speakers of the language or to a computer)
> > the student will naturally learn by himself. Later on, as we do in English
> > composition classes, we can teach him the refinements which are
> > undisputably valuable. As has been pointed out several times, learning yet
> > another language at this point is usually very easy.
>
> I'm not sure I agree here.  I had a student last year in a problem solving
> course using Quick Basic.  This student had (I'm convinced) been brain
> damaged
> by excessive exposure to the GOTO.  He wrote one of the most convoluted
> complex
> pieces of code I ever saw to do a rather simple and straight forward
> programming assignments.  Not only wasn't it reusable or extensible, it was
> unreadable!

Well Frank, I notice that you didn't say that it didn't work. From the
student's point of view that might be less important than pleasing the
teacher, but from the learner's point of view, if the code compiled and
ran, he might well have achieved his goal anyway.

Leigh

*If you know the history of Macintosh that will surprise you. No, I don't
mean Lisa; I mean Macintosh. It may surprise you to learn that the first
computing I did was done on a home computer in 1956. The home belonged to
Harry Huskey of the University of California Electrical Engineering
Department. The machine was his Bendix G-15, and you can see that very
machine on display in the Smithsonian (unless it has been removed in the
last three years).