Re: computer languages -- RPG400-L

saustad wrote:
>For my 2 cents worth, the best computer language ever designed is PL/I.
As
>for why it never caught on...

I agree - PL/I was always rather underrated.  It's also the source
for some of my favorite stories.  (Oh no!  Run away!  He's
starting on his stories again!)

In my 3rd year programming languages course, the prof discussed
the important programming languages at the time.  He covered each
one under a number of specific topics.  When we got to PL/I, he
wrote on the blackboard "Significant Enhancements", left the rest
of the board blank, and went on the to next topic!

The thing to remember about PL/I was that it was developed just
before programming language theory had settled down.  It was
before the benefits of strong typing and structured programming
were fully understood and accepted by most in the field.
Furthermore, the language was supposed to combine all the best
features of the common languages COBOL and FORTRAN.  And so the
language ended up with some interesting features.  The language
included pretty much every feature known to programming
languages at the time (including the kitchen sink).  Plus, the
language was designed with the naive and optimistic expectation
that the programmer knew what he was doing, and it was up to the
compiler to make sense of the code, however malformed.  (In a
way, PL/I was not only an ambitious programming language project,
it was also an ambitious AI project!)

Around the same time, another group was developing the language
(eventually called) Algol 68.  For a while, it appeared that the
committee would adopt a language designed by Niklaus Wirth.  But
late in the cycle, another ill-defined proposal was introduced,
which gained wide acceptance.  To make a long story slightly
less long, the committee adopted the new proposal, which had
lots of problems, and took a long time to develop, and ended up
with a totally incomprehensible description.  Implementations
of Algol 68 eventually appeared, but no one ever really took the
language seriously.  Wirth (and a few others) stormed out of one
meeting, and he continued to work on his own languages.  Algol W
morphed into Pascal, which then took the world by storm.

The point of that digression was to illustrate what was
happening in the world of programming language design at the
time.  Wirth argued in favor of strong typing, structured
programming, and small languages.  The Pascal language strongly
demonstrated his point, and became the favorite language in
academia.  And PL/I (remember PL/I?) was the total antitheses.

Okay, that was academe.  What about the business programming
world?  Well, PL/I was intended to replace other programming
languages like COBOL and FORTRAN.  But COBOL and FORTRAN
programmers had no desire to learn a new language, and so
instead of unifying the commercial programming world, it simply
further divided the programming community.

In my opinion, PL/I got an unfair rap.  Even before languages
like Pascal, there was little you could do in Pascal that you
couldn't also do in PL/I. It just required a bit of discipline.
Furthermore, PL/I subset languages (such as SP/k and PL/C) were
indeed being used to teach programming to university students.

But that's all ancient history now.  Now we have "object
oriented" languages and "scripting" languages to add into the
mix.  Now, the "cool" languages are Java, C++, Perl, and (my
favorite) Python.  Each has its own benefits and quirks, and
over time may also fall by the wayside.

Cheers!  Hans

Hans Boldt, ILE RPG Development, IBM Toronto Lab, boldt@ca.ibm.com