[Dev] PyCon trip report
Katie Capps Parlante
capps at osafoundation.org
Fri Mar 3 17:50:45 PST 2006
Thoughts about pycon this year...
I used Chandler 0.6.1 throughout the conference to look at the schedule
and pick sessions, after importing the pycon calendar. I also ran the
trunk, and gave demos of both throughout the conference. Several people
approached me about wanting to collaborate with us, but were not staying
for the sprints -- I hope to see some of them show up on the list. I
think I talked to even more people who were interested in *using*
Chandler than interested in hacking on Chandler. (I'm not sure what that
means). When I asked about why they wanted to use Chandler, they tended
to mention cross platform calendaring for a small group, especially if
it handled timezones and free-busy well. Several people mentioned
wanting Chandler to interoperate with exchange (presumably because of
their work environment).
My biggest obstacle to using chandler at the conference was the
performance (particularly on my mac). I'll write up my dogfood feedback
in more detail for the design list.
Though it wasn't technical, I found the first PyPy talk to be very
interesting. Their relationship to the consortium of organizations that
fund them, and the EU, reminded me of our relationship to Mellon and
CSG. They talked about the "complicated balance of interests" between an
open developer driven project and a funder that wants to track goals and
meet deadlines. To coordinate, they meet weekly on irc for 30 minutes.
They also sprint regularly, getting together in person every 6 weeks for
week long sprints. They noted that these core sprints weren't optimal
for the larger community, but very useful for the core team to
collaborate. They also do conference sprints, where the objective is to
grow the community.
People at the wxPython BOF were tackling a lot of the same problems that
we face, including printing and text/html editing. The projects at that
BOF envied our large team and the resources we are able to put into wx.
They all tend to work at the Python layer, and don't often venture into
widgets proper. The question of automating testing came up -- I think we
could do an interesting talk about our work in that area for some future
conference (PyCon or OSCON).
People at the Chandler BOF were very interested in tagging -- I think
they'll be jazzed to see 0.7. One person was impressed with the data hub
capabilities that we have given the twisted back-end -- he suggested
that we mention it more explicitly on the website/wiki/documentation.
By the time we got to the sprints, we were all sort of sleep deprived
and brain-fried. Jeffrey managed to have a good vobject sprint and we
had a good discussion about data/schema upgrades (both described
Reflecting on last year's sprint, I think this year's Chandler sprint
demonstrated that the schema API work paid off handsomely -- the parcel
written by Ralph, our sprinter, was written quickly and fairly
painlessly in comparison to last year. To learn more at the next sprint,
I think we need to have a different kind of exercise.
Grant and I worked on vtodo and vcard import, respectively. It shouldn't
be too hard to finish this up, once we get everything working with the
new and improved vobject. We do need some refactoring work to the
import/export framework (some of which is already logged in bugzilla)
before an external contributor could make use of it. I also have some
suggestions for improving the documentation based on the sprint (and the
earlier sprint in January) -- I'll add this to the 0.7 plan.
Notes from the conference and sprints here:
More information about the Dev