Cobol - Linugistics & COBOL design (was: Cobol AS/400 BNF Help Please

Author

Linugistics & COBOL design (was: Cobol AS/400 BNF Help Please

William M. Klein

2004-07-20, 3:55 am

As someone with a college major in COBOL (and who worked on the COBOL Standards
committee), I don't think there is any particular value in understanding how
"natural" languages work when designing (especially enhancing an existing)
programming or artificial language.

My memory (and I could be mistaken in this) is that Chuck (also from J4) either
had a degree or a minor in linguistics.

Having said that, I can easily see how my background and interest in
"theoretical" (rather than historical, etc) linguistics related to my initial
interest in programming.

--
Bill Klein
wmklein <at> ix.netcom.com
"Robert Wagner" <robert.deletethis@wagner.net> wrote in message
news:40fc7805.29951463@news.optonline.net...
> epc8@juno.com (E P Chandler) wrote:
>
> news:<nj4of0tsaulqu6u82e1o5qmikj2st9oopg@4ax.com>...
>
> That's because BNF strongly prefers languages that are 'context-free'. It
> believes lexicon should do ALL the heavy lifting. Its ability to handle syntax
> is very limited.
>
> This preference has influenced all languages designed after roughly 1970. It's
> the reason why C has '==' as a relational operator, for instance.
>
> <troll alert> It seems to me that 'language designers' should know something
> about linguistics. Knowledge of logic is inadequate. When challenged to name a
> famous linguist, most language designing computer scientists can name Chomsky
> and no others. When asked to name his seminal work, they cannot come up with
> Syntactic Structures. When asked about Bloomfield or Wittgenstein, they return
a
> blank look.
>
> "In the land of he blind, the one-eyed man is king."

Robert Wagner

2004-07-20, 8:55 am

You and Chuck work on the only major programming language where context
(semantics) determines meaning. Your interest in linguistics supports my point.

COBOL is renouned for being easy to read. The reason is it's similarity to a
natural language, English. Don't you think that has value?

"William M. Klein" <wmklein@nospam.netcom.com> wrote:

>As someone with a college major in COBOL (and who worked on the COBOL Standards
>committee), I don't think there is any particular value in understanding how
>"natural" languages work when designing (especially enhancing an existing)
>programming or artificial language.
>
>My memory (and I could be mistaken in this) is that Chuck (also from J4) either
>had a degree or a minor in linguistics.
>
>Having said that, I can easily see how my background and interest in
>"theoretical" (rather than historical, etc) linguistics related to my initial
>interest in programming.
>
>--
>Bill Klein
> wmklein <at> ix.netcom.com
>"Robert Wagner" <robert.deletethis@wagner.net> wrote in message
>news:40fc7805.29951463@news.optonline.net...
syntax[color=darkred]
It's[color=darkred]
a[color=darkred]
return[color=darkred]
>a
>
>

james8049

2004-07-20, 10:48 am

As far as I know the only programing language actually designed by a linguist is perl.

Perl has many of the problems of a natural language (particulary English) a dependence on context, infering the the subject and object of a clause from context etc. etc.

It also rolls off the keyboard, and, well written perl is quite readable.

FLAMEBAIT
Cobol on the other had I find problematic:

Badly written if statements can be totaly ambiguous.

A western reader's brain is hard wired to process periods in "firmware" - you simply do not "see" them, which is why nearly every language since has picked the semi-colon or something equaly big and obviuous as a statement terminator.

A sentence with lots of long words does not make good reading in English, and, a convulted Perform or Inspect with 30 character qualified field names make some COBOL programs harder to read then the small print in an insurance contract.

/FLAMEBAIT

Sponsored Links