(PHP or Python) Developing something like www.tribe.net

Michele Simionato michele.simionato at gmail.com
Thu Apr 28 09:29:25 EDT 2005


<talking about programming in the large and in the small>

Well, let me specialize my sentence

> CherryPy and Quixote are for programming in the small

in

> CherryPy and Quixote are good for programming in the small

meaning that these frameworks help you when programming in the small.
This statement does NOT imply that they get in your way when you
program
in the large. I just lack the experience in programming in the large
with
CherryPy and/or Quixote.

Also, let me specify what I mean by "programming in the large":
if a single person can grasp the application, and the application
can be implemented by a single team, then you are programming
in the small.

You are programming in the large only when you have many independent
team of developers and you have coordination problems. In this
situation
a component architecture is supposed to help (and maybe it does, I lack

the field experience to give an opinion), whereas when working in the
small a component architecture can just get in your way.

It is easy to evaluate a framework in the small: it is enough
to ask to self the question "how much time did it take to me to write
my first Web site in that framework starting from zero?". OTOH, to
evaluate a framework for programming in the large takes years of
experience and practice, and I will not hazard any opinion ;)

Still, I believe it is possible to have a frameworks which is
scalable both in the small and in the large. Look for instance
at the programming language spectrum: Java was intended to program
in the large and it is pretty bad when programming in the small;
Perl, on the other hand, was meant to program in the small and does
not scale at all in the large. However, Python works equally well
both in the large and in the small. So, I think there is no
contraddiction between large and small, in theory. But in practice
there is, so I use Zope at work and Quixote at home ;)


                Michele Simionato




More information about the Python-list mailing list