[Tracker-discuss] Creating a Jython tracker

["Martin v. Löwis"]

> But it is thin on reasons for not going with Trac.  Does anyone
> remember the particular problems with Trac?  

There are still AMK's notes out there:

http://wiki.python.org/moin/AndrewKuchling/TrackerNotes

Unfortunately, the trac demo was provided very late in the call for
trackers, so it wasn't studied as intensively as the others.

I remember that not being able to do custom queries sounded like
a severe limitation to me. You can probably fix that with a plugin

It's not clear to me whether Trac supports email interaction; we
believed that this is a must-have feature.

One specific concern I had is that all issues came out on a single
page always, which is just unacceptable if you have 1000+ open
issues (as we have now). I understand that this can be fixed with
a plugin.

In any case, you need some formal process to take a decision;
people will never agree on a tracker. So it's either a dictator
decision, a committee decision, or a community decision - use
whatever process works best for your community, and then implement
it.

In the case of Python, it was essential to the success of the
migration that people had to show-case their trackers *before*
a decision was made, as admin-type people already got attached
to getting this up-and-running. If you want to see a similar
dog-and-pony show for possible future Jython trackers, I could
still try to get the Roundup tracker up-and-running, even at
the risk that the community decides in favor of Trac afterwards.

In such a scenario, the Trac supporters would have get together
and volunteer some people to actually perform the necessary
setup work, so if Trac wins, you have already volunteers for
the first three months into the tracker's life (our experience
tells that most volunteers will run away within the first 6
months after getting this set-up and adopted). For Python, it's
currently just me that does any maintenance on the tracker.

Regards,
Martin