Re: Two Thousand and Eight

Snidely <Snidely.too@[EMAIL PROTECTED]
> wrote:

>On May 15, 6:43 pm, Greg Goss <go...@[EMAIL PROTECTED]
> wrote:
>> I'm not sure why I have some huge mental block on learning c.  The
>> underlying philosophy (other than "Shave every keystroke!" and
>> assignments buried within expressions) isn't that far off from PL/I.
>
>I would be suspicious of that claim.  In fact, I am suspicious of that
>claim.  My limited understanding of PL/1 is that it is more akin to
>Multics, the which K&R and gang were revolting against, then it is to
>Unix, which is where the philospophy of C comes from.  The idea of K&R
>is to do things quickly and lightly, with room for formalism later.
>Multics had much greater formallity (something later im****ted into
>Unix when security became a bigger issue), and PL/1 seems to have been
>especially heavy into formallity.
>
>PL/M kept some of the formallity, but with half the language thrown
>away as unnecessary, the formallity also became less burdensome.
>
>BTW, I am one of those annoying people who likes both formallity and
>lightness.  I also enjoyed using Ada (from the brown Grady Booth book
>-- I have the "starry night" book, too, but never got all the way into
>it).

The question would depend on the formal (heh) definition of formalism.
My impression is that PL/I falls more naturally into modularity than
c.  

I think of PL/I as being more related to the IBM architectures.  I've
worked on Multics, but only in Fortran, and didn't do anything
challenging at the OS level.  I've used PL/I heavily in DOS/VSE (later
renamed ESA) and in MVS (a rename of OS/VS) and lightly in VMS.  I've
used PL/M in a CPM clone (TurboDOS), but the deep design of the
program had already been done and I was just fixing bad practices
after the fact.

In googling, I find that I was confusing PL/M with PL/80.  The latter
was the Kildall (Digital Research) version.  PL/M might have been less
feature rich.

Wiki claims that one of the reasons for poor penetration of PL/I was
the demarcation of the programming world into scientific and business
programmers in the sixties and seventies.  As someone who started in
Fortran and whose first professional programming work was an OS
conversion and some minor enhancements of a huge map-drawing program
(Contour4), but most of whose life was spent in business programming.
And both of the businesses were engineering-oriented (railway and
smelter) rather than finance oriented, so the productivity issues of
PL/I won over Cobol.  A science background programming business stuff
in engineering-oriented companies seems to sidestep the emotional
issues that wiki describes.
-- 
Tomorrow is today already.
Greg Goss, 1989-01-27