SCOUG-Programming Mailing List Archives

<< Previous Message <<

Date:	Mon, 12 Jan 2004 23:26:51 PST8
From:	"Lynn H. Maxson" <lmaxson@pacbell.net >
Reply-To:	scoug-programming@scoug.com
To:	< "scoug-programming@scoug.com" > scoug-programming@scoug.com >
Subject:	SCOUG-Programming: QA equals testing, Part One:Detection

Content Type: text/plain

"This is just syntax, Lynn. You were proselytizing code which
other compilers could not generate. Take a look at the
generated code which PL/I actually creates from your
examples. You admire the compactness and readability of
PL/I, but the generated code isn't magical. ..."

Peter,

I never claimed that the generated code was magical. I don't
know why I would ever make that statement. You fault me
for submitting examples for which PL/I can generate code and
other languages cannot. Why would I not prefer a language
demonstrably without limits in terms of data types and
operands, with a simple and robust syntax, and essentially
universal in application? If you have to start somewhere, why
not pick the highest point in the food chain?

At one time the optimizing compiler had three options: opt=0
(normal optimization), opt=1 (space optimization), and opt=2
(time optimization). Later they dropped the space
optimization.

"...And is the compiler smart enough to realize it can set up an
internal field and simply add PI to it on each PRINT rather
than perform the multiplication? A human programmer will
see that instantly. I can code-spam you with innumerable
examples which can be easily optimized in ways a compiler
will likely miss if you're hesitant to agree with me ."

Oddly enough only human programmers write compilers.
Anything they "see instantly" they can specify. The compiler
like all other software and associated hardware can only
follow instructions. You can spam me with examples that
programmers can code outside a compiler but not within. I
suspect that involves zero writing.

As you can see I have no hesitancy on this matter.

Please, please, please never again ask me if a compiler or any
piece of software is "smart" enough to do anything. Never
ask me if computer hardware is "smart" enough to do
anything. In both instances the answer is "no". There is no
intelligence there. There is some intelligence needed in the
code writing. There is no intelligence transferred to or
present in the code written.

You've just offered one of the advantages of open source
relative to code generation in compilers or interpreters. As
you gain greater insight into code optimization you can
incorporate it within the code generation source. You can
also take advantage of those with greater insight than yours.

=====================================================

To unsubscribe from this list, send an email message
to "steward@scoug.com". In the body of the message,
put the command "unsubscribe scoug-programming".

For problems, contact the list owner at
"rollin@scoug.com".

=====================================================

Return to [ 12 | January | 2004 ]

SCOUG, Warp Expo West, and Warpfest are trademarks of the Southern California OS/2 User Group. OS/2, Workplace Shell, and IBM are registered trademarks of International Business Machines Corporation. All other trademarks remain the property of their respective owners.