I both agree, and disagree with you, James.
I have been an APL fan, and successful user since the late '70's. It
didn't
take me long at all to learn the character set. APL should have
become a
favorite, mainstream programming tool. Obviously it did not. One
reason as
already pointed out is that in its early days, it was ahead of existing
hardware, and only worked well only on mainframes - you know what they'd
cost in today's dollars!?! It could even cause an older HP
"minicomputer"
to dim all its lights! With the advent of IBM PC's and all newer
machines,
APL performed like a champ - so that impedement disappeared. It's
greater
productivity has always been known.
I sincerely believe that APL, and its successor, J, have both failed in
popularity for reasons having nothing to do with technical issues, but
rather with issues related to sales and marketing of the APL and J
products. In addition, it's hard to compete with C. With the full
backing of the old
Bell Labs, and the Unix world, that's quite a challenge! But the real
reason is not to be found on blaming stuff on others.
The late, great Dr. Ken Iverson after leaving academia spent many
years with
IBM, and, I suspect, was indoctrinated in the IBM ivory tower mentality,
which I, as a customer of IBM experienced in other areas altogether,
a real
turn-off. IBM is principally a hardware manufacturer, and not all that
oriented to software sales. Even when a number of IBM'ers left and went
into their own businesses of marketing APL, it was sort of like Kevin
Costner's "Field of Dreams" - if you build it, they will come! Baloney!
Just look at the majority of forum articles. I am a member of all the
forums except the Beta one. Many of the messages are oriented towards
getting the most bang out of one J sentence, terse, cryptic stuff.
Books on
APL warned against doing this because it leads to poor readibility,
and if
you're doing any serious development, you will have problems down the
road
maintaining this code. Those same articles reference the need of
extensive
commenting so that the code can be easily understood even by its
author even
six months later! Most NB.'s I see in these articles are little more
than
line titles. Before the days of the internet, such programming style
was
referred to as "cyberporn".
I agree that J is not a pretty language to look at, but you can say
the same
about C. So what? J extends the use of APL without losing its
productivity
and effectiveness, and that's what's important. I believe that the
use of
the ASCII character set was a wise choice because of the all the
ramifications of connectivity with the outside world.
I don't wish to appear to be pessimistic, but to avoid going in the
direction of the Sony Betamax, the purveyors of J must address the
issue of
sales and marketing. Also, they must really come to grips with the
GUI - a
fact of life in in the worlds of Windows, Mac, and Linux. Without
this, J
could be releagted to being a sophiscated desk calculator! Programs
today
are interactive. A real toolkit is needed, and it must be properly
documented. The functions are there, but more is needed. The J
purveyors
should like at professional quality C compilers and provide a similar
GUI
toolkit with similarly detailed tutorials. I have personal e-mails with
some of the key J purveyors, and while very polite and encouraging to
me,
they seem content with the current status of the J GUI for windows. I
respectfully disagree!
Most important, however, is for the J purveyors to get out there into
the
real world and push their outstanding product. It needs, and certainly
deserves this kind of effort to make it big time. Without that, it will
remain an academic curiosity. And the world will be poorer for it.
'Nuff said,
Bob from Boynton Beach, FL
