Difficulty Finding Python Developers

A B Carter gnosticray at aol.com
Fri Apr 16 09:46:18 CEST 2004


Peter Hansen <peter at engcorp.com> wrote in message news:<kKSdnSaWjO_rF-PdRVn-jg at powergate.ca>...
> A B Carter wrote:
> > 2 - Start setting expectations with management. For example, explain
> > that if there are not a lot of Python programmers it's because it's a
> > new language that
> > it builds on the strengths of langauges such as C++, Java and perl
> > while avoiding their weaknesses. 
> 
> If I were his management, here I might say, "But I thought you
> said Python was older than Java!  If it's so good, why hasn't it
> caught on more than Java?" or "How can it build on the strengths
> of Java when it's older?"
> 
First let me express embarrassment that I had no idea Python was as
old as it was. No doubt I'm biased by the fact that I've only begun to
use Python seriously this year. But damn, I still want to say that it
just seems really new and that yes, it builds on the strengths of
languages such as Java and Perl while avoiding their weaknesses.

So this changes the story you have to tell but certainly not the
reason your telling the story, which is to explain the lack of Python
programmers in a way that reveals a strength rather than a weakness.
You could talk about how Python was originally developed by a single
programmer for an in-house project which should have vanished in
obscurity except that it was so well designed that a kind of grass
roots campaign has made it the second most popular scripting language
in use. Think Guido=Linus and Python=Linux and so long as management
has read "Business Week", they'll be listening.

> > Mention that when you do find a
> > Python programmer he'll probably be better than your average perl or
> > Java programmer. 
> 
> And here I would ask, "Why do you say that?  If there are so
> few of them around, how could anyone know whether they're better
> or worse?  And wouldn't the best programmers be using the most
> popular languages, because they would make more money that way?"
> 
>  > Make the argument that you get what you pay for, and
> > the extra expense of Python is worth it.
> 
OK, this was contentious even for comp.lang.python. To this group I
can sum up the argument by saying the "Learn Python in 21 Days" and
"Python for Dummies" books are much better written than the ones for
Java and other languages. What you can tell management is that Python
is not an over-hyped language like Java (and yes if the manager was
once a Java developer then mention .NET instead) and that Python
programmers are people who really love the field and discovered Python
because of their dissatisfaction with other languages. If management
is willing to do a bit of reading have them read Eric Raymond's essay
on why he switched from Perl to Python.


> "I thought you said Python was free... now you're telling me it's
> going to cost us more than the other languages because it's so hard
> to hire anyone who knows it?"
> 
Sorry, but this I don't buy. If management thought they were getting
something for free then both parties are at fault. The choice of any
language is a matter of balancing trade-offs and management should
have been informed of both the benefits and the risks; and if
management didn't ask about the risks then they just weren't doing
their job.

> > 3 - With 2 out of the way consider rethinking how this three month
> > project should be done. If Python talent is scarce then it might make
> > more sense to develop in-house talent. This may no longer be a three
> > month project but down the road your in a more solid position, which
> > ,if you think about it, is a basic part of what Python is all about.
> 
> Good advice here, IMHO. :-)  This is what we did when first considering
> Python, by the way, and I had a co-op student learn the langauge and
> develop a GPIB interface with a wrapper around a DLL using "calldll".
> It took him about a week to learn Python, about a week to get a basic
> calldll wrapper working, about a week to implement a simple RF test
> for one of the devices the company made (controlling two signal
> generators and a spectrum analyzer) and about a week to figure out how
> to slap a simple GUI on it using Tkinter.
> 
Thank you, but if you accept point 3 then something along the lines of
point 2 is an absolute must because you've gone from asking money to
pay for three months of consulting to hiring a full-time programmer.
Strictly in terms of dollars spent that is, by my accounting, roughly
five times more expensive. You got some explaining to do.

Regards, A B



More information about the Python-list mailing list