Re: The joy of FORTRAN

On 27/02/2025 09:30, c186282 wrote:

On 2/27/25 2:43 AM, Charlie Gibbs wrote:

On 2025-02-26, Peter Flass <peter_flass@yahoo.com> wrote:
>

c186282 <c186282@nnada.net> wrote:
>

On 2/25/25 2:33 PM, vjp2.at@at.BioStrategist.dot.dot.com wrote:
>

I went from DEC20 FORTRAN to pdp11 Basic (it had matrix math) to DEC SAIL
(Algol) to Pascal to C to Python

>
   Sounds familiar ... though 'C' became available on
   the PDP-11s (was writ on them). Still pref Pascal
   over Python where possible.
>

Assembler and COBOL were needed but avoided

>
   Better ASM than COBOL  :-)

>
Depends on what you want to do. Assembler is a lot more fun,
but I wouldn’t want to write a payroll system in it.

>
Wrote 'em, maintained 'em.  They're hell in any language.
>
Having said that, I really enjoyed working in assembly language,
for the privilege of not having some snooty compiler slap your
wrist and say you needed a page of code to do something you could
do in a few lines of assembly code.
>

   ASM can give you kind of a buzz, makes you one
   with the machine.
>
   Alas if I'd learned more COBOL then I could have
   had a lucrative retirement income supp maintaining
   all those old biz/ops code. Still LOTS of it in
   use and it's too expensive now to replace. If it
   works you hang on to it with a death grip.

>
It’s not just COBOL any more, it’s all COBOL/CICS/DB2. Same with PL/I.

>
My COBOL days didn't involve any database stuff.  Mind you, I've
managed to avoid any sort of DBMS for my entire career.  However,
I did do enough work with Univac's equivalent of CICS (in both
COBOL and assembly language) to be glad to be done with it.

    We ADMIT - COBOL SUCKED.
    The "All purpose business language" was just
   HORRIBLE to work with.
 

That is probably because you don't have the patience to stick to the rules.
What mattered in COBOL applications was no bugs. None. Not one.
The process of analysing a business and creating business oriented applications that were 100% reliable lead to system analysts working out how to do it and passing the specification down to programmers and testing every single module.
Worth it for banks and big companies.

   I've done some COBOL - but under 100 lines.
   Did NOT love it.
 

I looked into it quiet hard at one point. I was very impressed. It was clearly a language that a total numpty could program in and get right.
When I went into databases etc I found some of the methodology to be useful as well.
Also it was clearly designed for machines with limited memory with modules not being held in RAM if not in use.

   On the flip, a HUGE volume of biz/gov stuff
   that STILL WORKS was writ in COBOL during
   the 60s. It was good, it worked, it STILL
   works. Those 60s nerds were GOOD.
 

I think it was less that the nerds were good than that they followed strict guidelines of how to code.
At some level everything was flowcharted and every single bit of data memory had a name and its function went into the data dictionary.
In short it was fully documented before it was written, in detail, and then passed over to hack 'progs' to turn into COBOL
And TESTED against the detail specifications
This is big corporation stuff .

   Oh, for Linux, some COBOL development
   environments - including 'IDE' - are to
   be had. It's NOT a dead language - just
   'less popular' than it used to be.
 

IIRC Microfocus COBOL still makes a profit.

   If you ARE a COBOL expert - there's LOTS
   of money to be made. Replacing that COBOL
   is now TOO EXPENSIVE ... so maint has become
   a Big Thing.

IIRC the main money was in RPG?
i.e. once the big cobol program was actually written and you had the data, you needed tools like RPG to produce decent reports and they always wanted more
--
Microsoft : the best reason to go to Linux that ever existed.