py3k s***s

Donn Cave donn at u.washington.edu
Tue Apr 15 13:09:15 EDT 2008


In article 
<11567a1a-b184-42e6-bbf3-a655736c19c7 at p25g2000hsf.googlegroups.com>,
 Sverker Nilsson <sn at sncs.se> wrote:

> No one forces me, but sooner or later they will want a Python 3.0 and
> then a 3.1 whatever.
> 
> I don't want that fuzz. As about the C versions, I am not that
> worried. What's your point?
> 
> I just like want to write a program that will stay working. And maybe
> I can go on with something else hopefully than just compatibility
> fixes. They take some work afterall.
> 
> It seems hard with Python. Esp. 2 -> 3

Welcome to the world of Python.

There was a period of relative stability in the '90s, culminating
with version 1.5.2, which just happens to be about the time that
people started taking Python seriously.  It turns out that this
stability was only due to lack of resources, though.

I think most of us are working under two imperatives here that
really boil down to the same thing:   we want a programming
language that lets us focus on the goal of the program.  On
one hand, that means things like automatic memory management
that are brought to us by newer languages (i.e., less than
30 years old.)  On the other hand it means stability that allows
our deployed code to go on working without constant intervention,
which usually we find in languages that have become utterly boring
and out of fashion (i.e., more than 30 years old.)

It's hard to find a good compromise between these two, in an
interpreted language.  I don't know what the current party line
may be on this matter, but some years back it was that you should
consider the interpreter part of your application.  That is, each
application should deploy with its own dedicated Python interpreter,
complete with libraries and everything.  This naturally relieves
some of the maintenance issues, since at least you can upgrade on
your own schedule, but of course it has its costs too.  Anyone who
might be thinking about using Python for an application should
seriously think about this.

   Donn Cave, donn at u.washington.edu



More information about the Python-list mailing list