🚀 go-pugleaf

Burton has agreed to my statement that ``computer science and
programming have about as much to do with each other as mathematics and
programming or mathematics and physics.'' I gave him a chance to retract
his agreement, and he didn't. I conclude that he no longer supports his
previous contentions that a Department of Programming would be a mere
subset of the computer science department. So this argument is over. I
once again claim that the lack of a Department of Programming implies
that we know neither how to teach nor to test programming ability in any
``objective'' fashion the way that we can in engineering or medicine. In
other words, programmer licensing is stupid, because the licensing test
will necessarily be so superficial that it will have no meaning in the
real world.

However, John did make some factual errors in his most recent article,
and I would feel guilty letting those errors go by without notice.

In article <1991Jul29.151510.17927@news.larc.nasa.gov> burton@asdsun.larc.nasa.gov (John Burton) writes:
> good starting place for understanding
> random number generation, but that can't be...that involves number
> theory and manipulation of discrete numbers...thats not programming,
> that has NOTHING to do with programming, that is computer science...

Actually, most of the theory Knuth uses in that chapter is elementary
number theory, which is mathematics, not computer science.

> but then, what is it doing in Knuth's book??? (see CACM article
> "Random Number Generators, good ones are hard to find")

Here you imply that the referenced CACM article shares some quality,
either content or at least technical merit, with Knuth's discussion of
random number generation. As a matter of fact, the article makes several
recommendations which are in direct contradiction to Knuth's statements,
supplies a generator with numerical qualities barely better than those
of RANDU, shows a distinct lack of technical sophistication, and is
usually dismissed by competent practitioners as an example of what not
to do.

> Many engineers started out as I mentioned, taking things
> apart and putting them back together...many graduated from *that* to
> do the typical teenage thing...modifying cars for better performance.

Once again you imply that the skill of a mechanic is connected with the
skill of a mechanical engineer. That is unsupportable. Most mechanics
show absolutely no interest in mechanical engineering.

> >                   My questions about hobbyists still stand, and you
> >know it. In most artistic endeavors---including programming!---there are
> >many recognized cases of young geniuses. I defy you to show that the
> >same is true of licensed professions.
> I just did...

No, you did not. There are no documented cases of a young genius in
mechanical engineering, or civil engineering, or even architecture.

> >> could you name some irrelevant material that is taught in computer
> >> science that would NEVER be used by a programmer in the real world?
> >Turing machines.
> Oh, I see the problem now...you're NOT a computer scientist. You come
> across the idea of a turing machine, and you're background is not
> strong enough for you to realize that a turing machine is the
> theoretical basis for PROGRAMMING languages...

Actually, son, I read and understood Godel's proofs before I even heard
the term ``Turing machine.'' I, unlike you, was trained in mathematics,
and so far in my career in mathematics have received more awards than
you will ever know. Several programming languages were designed before
Turing wrote his first paper. No language actually used for programming
is based in any way on Turing machines. They are at best a convenient
abstraction, and for programmers in the real world they are completely
irrelevant.

> You see a strict division between art and science, where none exists.

I have made no such statement. I see a division between art and
engineering. Programming is an art, and as such cannot be objectively
tested. I am not arguing the point that programming errors occasionally
claim lives. My claim is that we can't help the situation by licensing,
because we do not know a suitable test for the ability to write a
program. An engineer who can solve several dozen small problems and
tests of his knowledge will quite likely be able to apply that knowledge
to a larger problem. The same cannot be said of programmers.

> You've been very adamant about programming not being part of computer
> science and that schools do not teach *programming*...tell me, just 
> what *would* you teach in your "Department of Programming" ?

I have already told you this. The best way anyone knows to teach an
artist is to show examples of great (and awful) works of art from the
past, explain why they are great, analyze the ever-improving works of
the artist himself, and meanwhile make sure that the necessary mundane
skills (e.g., sweeping a brush across a canvas, reading and writing
sheet music, reading and writing a programming language) are adequately
developed. Sure, there's theory on small and large scales, just as there
is in music or in painting, but you can't simply ask someone to memorize
the theory, apply it to a few small problems, and expect that he'll then
be able to create beautiful work of his own.

---Dan