Moratorium on Python language changes?
Guido’s proposing a moratorium on changes to the Python language, the idea being to enable alternative implementations to have a stable target to aim at. This only applies to the syntax of the language, not to the stdlib, nor to the underlying implementation: the announcement hints at acceptance for a GIL-free implementation if one came along.
I’m all for it, myself. Like other people, I believe that the stdlib needs lots of care and attention in several respects. Removing one focus of changes will result in at least some of that effort shifting to the stdlib. (Altho’ it will probably result in some of the effort simply going elsewhere :) ).
j_king said,
Wrote on October 22, 2009 @ 2:57 pm
I’ve been told that having an implementation be the standard for a language is stupid. I tend to agree; it makes the language a moving target for alternate implementations and restricts a lot of what they could do.
Standardize the syntax and let the developers create better implementations.
Michael Dillon said,
Wrote on October 22, 2009 @ 6:04 pm
Hopefully some of that effort that goes “elsewhere” will go into creating a second tier Python toolbox, stuff that is not in the standard library, and probably will never be, but which we consider to be some form of best practice. This could be things like Beautiful Soup, or redis, or M2crypto.
Some of these third party projects have difficulty keeping up with newer CPython versions, and even where the module does not make sense on Jython or IronPython, it would be good to have a standard shim to access native Java or .NET libraries through the same interface. This would improve the portability of Python applications.