Not enough Python library development [was PEP scepticism]

[David Goodger]

on 2001-06-29 12:18 PM, Guido van Rossum (guido at python.org) wrote:
> But what exactly is the point of adding it to the
> standard library when you can get mxDateTime if you need it?

Several points! (Impatient? See #6.)

If a module is part of the standard distribution:

1. It gets wide exposure, hopefully resulting in bugs being found out and
improvements being made.

2. Bugs reports and patches can be contributed to a much larger developer
community than any solo module enjoys.

3. There's a greater chance of continuity if/when the module is abandoned by
the original author.

4. There should be a regression test for the module. It therefore has a
better chance of being kept up-to-date with the latest language features
(i.e., the module won't break when feature X changes, is added, or is
deprecated).

5. The module tends to have decent documentation, and that documentation is
placed in a central location and indexed.

6. (perhaps most importantly:) C extension modules get compiled for every
platform with a binary Python distribution. Not all of us have a C compiler
at hand, want to use one, or have the knowledge & experience necessary to
overcome code porting issues. (I used to know my way around a C compiler,
but I've been spoiled by Python! :-) For example, I don't see any Mac builds
of mxDateTime 2.0 on the net.

I think "Batteries Included" is a great goal. Adding to and improving the
standard library should be a goal of every release of Python. The more
quality batteries the better!

-- 
David Goodger    dgoodger at bigfoot.com    Open-source projects:
 - Python Docstring Processing System: http://docstring.sf.net
 - reStructuredText: http://structuredtext.sf.net
 - The Go Tools Project: http://gotools.sf.net