On Mar 5, 2006, at 22:10, Barry Warsaw wrote: > I'm happy to announce the release of the email 4.0a2 standalone > package. > This is the latest version of the email package and will be released > with Python 2.5. The major changes between this version and email 3.0 > (released with Python 2.4) is: ...
> Note that the old, email version 3 package names are still > supported for > backward compatibility, so you won't have to change existing code. > New > code should use the new names as the old names will go away in Python > 2.6. Also note that email.mime.application is /not/ provided as > email.MIMEApplication. I am a bit worried by annoucements such as this one, so I would like to profit from this opportunity to plead for more stability in Python development, even though the concrete case (the email module) is not a problem for me. My interpretation of the above paragraph is that it will be impossible to write Python code using the email module (and possibly other evolving modules) that works with both Python 2.4 and Python 2.6. As an author of Python code, I therefore have to choose if I want to force my users to update their Python installation (knowing that this is impossible or at least a major effort for some of them) or if I want to prevent early adopters of new versions from using my code. I don't like either option. From exchanges with users of my code, I conclude that there are significant numbers of users of Python 2.2, 2.3, and 2.4 at the moment, plus a small number of users of 2.0 and 2.1. From a few cases of Python 2.2 users whose situations I know well, I understand that updating is not always an option. For example, there is someone whose 10 PCs are still running SuSE Linux 8.1 (which includes Python 2.2). The PhD student who handles system administration (with minimal expertise) adopts the wise approach of never breaking a running system. In other words, those machines will remain at SuSE 8.1 and Python 2.2 until they will be replaced. This leaves users with the only option of installing a newer Python in their home directories - not exactly a straightforward task for someone who knows little about Python and nothing about system administration. Given the current situation, I expect that when Python 2.6 is released, Python 2.4 will still be in common use. Personally, I would prefer that all compatibility-breaking changes be postponed until Python 3K. I'd rather have worry about compatibility once than continously. Konrad. -- --------------------------------------------------------------------- Konrad Hinsen Laboratoire Léon Brillouin, CEA Saclay, 91191 Gif-sur-Yvette Cedex, France Tel.: +33-1 69 08 79 25 Fax: +33-1 69 08 82 61 E-Mail: [EMAIL PROTECTED] --------------------------------------------------------------------- -- http://mail.python.org/mailman/listinfo/python-list
