Upgrade 2.6 to 3.0

Steve Holden steve at holdenweb.com
Thu Feb 5 04:15:11 CET 2009

Tim Rowe wrote:
> 2009/2/5 Giampaolo Rodola' <gnewsg at gmail.com>:
>> Just out of curiosity, am I the only one who think that switching to
>> 3.x right now is not a good idea?
> I'm looking at making the switch, but I'm put off by the lack of 3rd
> party stuff such as PyWin (and I can't see a NumPy build for Python
> 2.6 yet, never mind 3.0). Unless all you want is in the standard
> library, I think it's worth the general user holding back for a while
> whilst the tool providers catch up.
Take a look at the recent threads on the topic of the 3.0.1 release on
the python-dev list. There is some feeling that the 3.0 release was
unsatisfactory in certain ways. Some of the issues (including removal of
some obscure features that should have been removed in 3.0) will be
taken care of in 3.0.1.

In addition to that expect a 3.1 release before too long (say, within
the next six months) that will address some performance issues not
addressed in 3.0.1 and provide other optimizations and perhaps a small
list of new features. 3.0 is of remarkably good quality for a ".0"
release. Expect even better things in the future.

If you want to write portable code you can put into production now and
move to 3.0 when appropriate then write for 2.6, and 2.7 when and if it
comes out. Compile it with the -3 option and rewrite until it stops
raising warnings. It will then be relatively easy to move to 3.x when
the third-party library support you need is available.

Steve Holden        +1 571 484 6266   +1 800 494 3119
Holden Web LLC              http://www.holdenweb.com/

More information about the Python-list mailing list