Choosing a programming language as a competitive tool

[Steve Holden]

"Courageous" <jkraska1 at san.rr.com> wrote in message
news:tq70ft0r480sb2mh94g050qoc2airk96t9 at 4ax.com...
> On 02 May 2001 10:44:04 -0400, Andrew Kuchling
<akuchlin at mems-exchange.org> wrote:
>
> >John Schmitt <jschmitt at vmlabs.com> writes:
> >> What I got out of this was that choosing a good programming language
can be
> >> a competitive advantage.  Graham chooses Lisp.  For me this isn't a
Lisp
> >
> >I liked all the comments on Slashdot that said they'd prefer to use
> >something more common such as C++ because it's too difficult to find
> >good programmers who can handle Lisp.  Hmmm.... "let's not get good
> >programmers to make our product in Lisp/Python/whatever; we'll get
> >some mediocre programmers to make it in C++ and that'll be better for
> >the company."  Explains a lot about the pathetic state of software,
> >doesn't it?
>
> Not really, no. Not only is it difficult to find Lisp programmers, it's
hard
> to train them, and even harder to make them good Lisp programmers.
> Add to this mix that there's little out there in terms of third party (pay
or
> free) library support, and you have a really sour recipe. In my
experience,
> if you're going the Lisp route, you need mavericks.
>
The converse side if this argument, of course, is that if you pay peanuts
you get monkeys. Monkeys are in plentiful supply, are cheaper than
mavericks, and won't complain about obviously stupid policies like "All
knowledge-based systems using advanced AI techniques will be programmed in
VBScript". If you're going the LISP route you need intelligent and motivated
staff, who will probably command a premium salary.

> Python, however, is different. Whether they know the language or not
> doesn't matter much. They can learn all they need to know in under a
> week, and the library core, both included and supplementary, is
> enormous. There's also much higher support in the free community
> (the net, etc).
>
Unfortunately a monkey writing Python is still a monkey. I agree that Python
is well-suited to many net-oriented tasks, and can be learned reasonably
quickly. But there's no substitute for experience, and an inexperienced
programmer will, unless carefully supervised, write bad programs in whatever
language they are told to use.

The main problem underlying the lack of good programmers is most
organizations' unwillingness to invest in training their staff. They would
rather poach from others than homegrow their programming talent, and never
bother to think what kind of mindset this engenders in programmers. And yet,
dammit, programming work is so much more fun when you have others around you
who will help you to learn and grow. In such an environment *everybody*
learns to be more productive.

There's a kind of Peter principle in operation, dictating that if you're
good enough at programming you will eventually end up in charge of a bunch
of monkeys, with nobody who can train you.

Disclaimer: I make a living from training, so may be biased.

and-that's-why-i-work-for-myself-ly y'rs  - steve