[Python-3000] pre-PEP: Procedure for PEPs with Backwards-Incompatible Changes

Guido van Rossum guido at python.org
Tue Mar 28 20:55:59 CEST 2006


On 3/28/06, Ian Bicking <ianb at colorstudy.com> wrote:
> skip at pobox.com wrote:
> >     >> Suggestion: Make this PEP 3001 and start any Py3k PEPs from 3100....
> >
> >     Guido> I already proposed that numbering scheme. More formally, Py3k
> >     Guido> meta PEPs go between 3001 and 3099, and feature PEPs start at
> >     Guido> 3100 (and hopefully we won't have to overflow into 4000 :-).
> >
> > Should there be some distinction between Py3k PEPs which fall under the
> > purview of Steven's PEP and those which contain completely new stuff and
> > aren't going to impact Python 2.x code?
>
> I notice also that some of these suggestions are applicable to 2.x, like
> the "Parallel iteration syntax" (which introduces no backward
> incompatibilities).  If something can be applied to 2.x, should that be
> brought up in py-dev instead (or c.l.p.)?
>
> There seems to be a danger that Py3K is seen as a more friendly place to
> suggest changes than Python 2.x (or maybe that the python-3000 list is
> more friendly to these suggestions than py-dev), and so changes are
> brought up here even though they could be applied earlier.  This would
> be problematic in part because if a change *can* go in 2.x, it would be
> really good if it did, so that the move to 3.0 involves as few changes
> as possible.
>
> Formalizing the target implementation through the PEP numbering might
> also cause premature expectations about when the feature might be
> introduced.  Though if it's okay to just renumber the PEP then that
> wouldn't be a problem.

Right. I guess this is one of the meta-issues that we should decide
upon shortly. I like your strawman: if incompatibilities or synergy
don't require it to go into Py3k, let's propose it for 2.x. (But in
general it's too late for 2.5, unless you have a *really* small tweak
or a *really* important issue. We don't want to make 2.5 late.)

--
--Guido van Rossum (home page: http://www.python.org/~guido/)


More information about the Python-3000 mailing list