Re: [Nolug] Evo and Outlook

On Monday, July 14, 2003, at 04:07 PM, Ron Johnson wrote:

> On Mon, 2003-07-14 at 09:43, Judson Lester wrote:
>> On Friday, July 11, 2003, at 06:11 PM, Ron Johnson wrote:
>>
>>> On Fri, 2003-07-11 at 14:43, rbstrickland@cox.net wrote:
>>>>> Thanks. I wonder how many of these vulnerabilities would be elim-
>>>>> inated if most programmers stopped using C, and went to pointerless
>>>>> languages (everything from Python, BASIC, Java and Lisp/workalikes,
>>>>> all the way to COBOL, Ada and FORTRAN) for most work.
>>>>
>>>> I understand the point you're making, but COBOL does have pointers,
>>>
>>> Yes, COBOL-99 does, but no one really uses it. COBOL-85 is *the*
>>> 3GL business language.
>>
>> Blah. COBOL is a ledger book with an attitude.
>
> Bull.
>
> It is a language tuned to a record-oriented paradigm (and that
> means records in all their forms, including embedded variable
> length arrays), but so what? Many things in commerce have that
> orientation. That's why COBOL-85 is so damned useful.
>
> When I was in college, the way I learned COBOL-74 made everyone
> in the classroom *despise* it, because it made us go through
> such incredible contortions to avoid GOTO (Dijkstra's evil
> legacy) in a language that doesn't have ENDIF or DO-WHILE.

Okay, I think I just found out that I'm the far left as far as
programming paradigms go here. It seems really bizarre to see another
proponent of self-documenting code call "goto considered harmful"
"Dijkstra's evil legacy."

I'll admit my COBOL experience is limited to a casual flirting, but in
that I can to two conclusions, which I'd be very pleased to be
disabused of. One, COBOL is strongly oriented around decimal records,
making it very limited outside of its intended realm, being (financial)
business logic. (A subconclusion was that that realm was limited
enough that it could be pleasantly implemented in a library of another
language.) Two, that the control structures of COBOL were limited
enough that it would be nearly impossible to code anything worthwhile
with producing spaghetti code with goto-sauce.

Those two understandings, combined with the degree to which data/code
mixing is encouraged lead me to the "ledger book with an attitude"
opinion.

Judson

___________________
Nolug mailing list
nolug@nolug.org
Received on 07/15/03