Efficacy of code generators

[Talking about code generators/CASE tools]

>The cost of these proprietary languages is
>relatively high.  The performance of the
>executables is relatively poor.  Programmer
>productivity is about the same.

It's been an interesting thread.  I wonder about the omission of code
quality.  Certainly, the average Synon program doesn't have as many careless
mistakes as the average hand-coded program.  Robust applications are better
than fragile ones I think.

Our Synon code rarely (never?) throws invalid array index errors or divide
by zero or any of a host of unpleasant things that the creative hand-coder
somehow never manages to clean up, being more concerned with the relative
merits of MOVE vs. Z-ADD than with solving a customer's business problem
with a robust, quality application.

Do I think that it is impossible for hand-coders to do a good job?  Not at
all!  I _do_ think it is more rare than we believe, and that CASE tools
raise the overall quality level of the application suite.

In this day and age, we don't write a 3 program application anymore.
Requirements are more complex, resulting in more complex solutions.  From my
limited experience, far too many "creative" programmers are loose cannons in
the complex modern environment.  It's simply not possible to "know" all of
the source code in a modern application, but how often have we heard
somebody complain that they can't follow the code because of those darned
"/COPY" statements or subroutines or subprocedures.

CASE tools ARE efficient because they let an average programmer produce
above average quality in her application.

"Creativity is like driving a car at night.
 You never see further than your headlights,
 but you can make the whole trip that way." -- E.L. Doctorow
  --buck