[Kevin]
> Also, any chance of a quick post-PyCon report at the next meeting from
> one of the pilgrims freshly returned from Chicago?

This is really something better discussed in person, but very quickly

1. The desire to switch to DVCS is driven by a need for a better way
for the developers of the various language implementations (cpython,
jython, ironpython, pypy) to collaborate, particularly on the standard
libary.

2. Some jython committers (Frank, Jim B, and Phil J) now have commit
rights to the cpython repo, so that fixes from jython can be pushed
upstream.

3. AFAIK, pypy folks already had those commit rights.

4. Ironpython folks have legal problems with pushing stuff upstream,
presumably because they would need publish modifications to PSF-owned
code that is distributed with ironpython (which is under an MS open
source license).

5. Unladen swallow was a point of major interest, because it has
already achieved a 30% speedup of cpython, and hopes to eventually
achieve a speedup of 500%(!). It is an open source project, and there
is a full intention to make all improvements available to cpython. The
30%-speeded-up version is running youtube, as of last Monday.

6. But the other implementations are also catching up. Jython is
getting big speed improvements, thanks to the advanced compiler work
of Tobias Ivarsson.

http://twitter.com/thobe

7. Ironpython is still experimenting with new compilation techniques

http://ironpython-urls.blogspot.com/2009/03/ironpython-26alpha-1-released.html

8. One of the most impressive projects presented was Adrian Holovaty's
everyblock.com (video not yet available, watch out for it here:
http://pycon.blogspot.com/). Check out everyblock, it's an amazing and
slightly scary (from privacy-elimination POV) project, to be
open-sourced on June 30th when their current round of funding expires:
Adrian is keen for everyone to use the code in their local city: at
least 50 people in the audience put up their hand when he asked who
was interested in doing so). Everyblock currently covers 11 US cities,
with more to come.

9. Python 3.0 is effectively 'end-of-life'd: users wanting Py3K should
move straight to 3.1. There may or not be python 2.x releases after
2.7, depending on how quickly the community moves to py3k.

10. There was much talk of a better packaging and distribution
process, to fix the problems of disutuils and setuptools. Tarek Ziade
is the lead on this effort (http://tarekziade.wordpress.com/), writing
PEPs, etc. VirtualEnv and zc.buildout both address problems in this
space.

11. Several meetings were had to try to address problems with WSGI

http://mail.python.org/pipermail/web-sig/2009-March/003693.html

But thanks to the existing (somewhat flawed) WSGI, cpython frameworks
are now running in more places

http://docs.djangoproject.com/en/dev/howto/jython/
http://pylonshq.com/docs/en/0.9.7/jython/
http://lackingcredibility.blogspot.com/2009/03/twisted-django-it-wont-burn-down-your.html

12. Endless other stuff that I'm not going into here.

I wouldn't mind meeting up on Wednesday (i.e. a week before the next
meetup) for a drink, to report on this stuff while it's all still
fresh.

Questions on specific areas might be easier to answer.

Alan.

--~--~---------~--~----~------------~-------~--~----~
You received this message because you are subscribed to the Google Groups 
"Python Ireland" group.
To post to this group, send email to [email protected]
To unsubscribe from this group, send email to 
[email protected]
For more options, visit this group at 
http://groups.google.ie/group/pythonireland?hl=en
-~----------~----~----~----~------~----~------~--~---

Reply via email to