PEP scepticism

Barry A. Warsaw barry at digicool.com
Fri Jun 29 06:26:14 CEST 2001


>>>>> "PP" == Paul Prescod <paulp at ActiveState.com> writes:

    PP> It annoys me when people equate the PEP process with the
    PP> adding of features. Features were added before there was a PEP
    PP> process. Features have been added with out a PEP even since
    PP> the process existed. The PEP process also handles things like
    PP> library upgrades and distutils improvements. If you think
    PP> Python is changing to quickly, that's fine.  But what does it
    PP> have to do with PEPs?

The other purpose for PEPs -- IMO more important and often overlooked
-- is to provide a concrete historical record for decisions made about
the language, its implementations, and its environments.  Python's
been around a long time, and as Tim is fond of saying, there are
almost no new feature suggestions that haven't been brought up many
times before.  The PEP is now the place which records the decision
making process and in that, it serves a very useful function, even
(and it some cases more) for when the idea is rejected as when it's
accepted.

That's quite apart from how the PEP process has influenced the pace of
changes to Python.  My own personal opinion is that the pace hasn't
really changed much over time.  What /has/ changed is the size and
scope of Python's user base, and the relatively smaller advancement in
elegance or power that current changes make, compared to what you can
already do in the language.

That's not to say that Andrew doesn't make a very important point.  In
one concrete example though, I know that with my own attempt at
improving Python's MIME offerings, I've found that features like
extended print have definitely helped the elegance of the library.  If
I was willing to target Python 2.2 then iterators and generators would
as well (I think ;).

I have a nagging feeling that `creeping featurism' is like pork-barrel
politics.  In one breath people will decry the special interests that
they don't like or don't care about, while in the next, ask for their
own pet project to be funded.  It was quite funny to see the same
effect at work at IPC9.  Lots of hands went up when Guido took a poll
on whether the pace of change was too fast.  But as the session
continued, there were certainly no shortages of new features
suggested!

-Barry




More information about the Python-list mailing list