Productivity and the two language approach

[Derek T. Jones]

> Seriously, Python takes all of ten minutes to begin coding in, all of an
> hour to link into your C code (can't comment on c++, anyone else?), ...

I've linked C++ code into Python, and it took less than an hour.  The only
thing different is that you have to surround the "init<module>" function
with the 'extern "C" { }' business.  That's all I noticed, at least on
Solaris.  Simple.

To put another $0.02 in:  I am an expert C++ programmer with > 4 years
experience with it by now.  Learning Python was extremely easy, far more so
than Perl, which I have never picked up all the way.  One reason for this,
I think, is that Python's approach is more like C++, so to me, it's
familiar.

I share the original author's skepticism about having to "learn new
environments", but it only applies when the new environment is very
dissimilar from the one you're using, difficult to learn, or both.
Based on my own experience, I would strongly encourage the two-language
approach in this case.

Another spontaneous Python testimonial:  last Friday I solved a text-
processing problem for someone in the amount of time it took to describe
the problem.  My "customer" was amazed.  He'd thought we were just writing
psuedocode in preparation for the solution itself -- suddenly, answers
were coming out.  Not just that, but he had a reusable script.

I love C/C++ for big, performance-critical projects, and won't ever give it
up, but Python beats C++ for top-level integration, and rapid prototyping.
Hands down.

--Derek

-- 
My real email address has only one 'n' in it, for those who care.