Re: [Pyblosxom-devel] issue tracking -- let's fix this!

[Marius G. <ma...@ge...>]

On Thu, Sep 30, 2010 at 03:14:50PM -0400, will kahn-greene wrote:
> I really want to solve number 1 next.  An issue is a bug, feature,
> plugin idea, documentation suggestion, ...  I want to use an issue
> tracking system that is:
> 
> 1. maximizes ease of use
> 
> 2. minimizes spam
> 
> 3. maximizes ability of anonymous people to use -- people shouldn't have
> to create an account to log a bug (maybe this is solved by people
> sending me email--I don't know)
> 
> 4. maximizes the ability to figure out where PyBlosxom is at in terms of
> issues, items that should be done, things that people can jump in and
> work on, ...
> 
> 5. minimizes administration, maintenance and installation difficulties
> 
> 6. must be Free Software

1, 2 and 5 are best solved by using a hosted bug tracker (Launchpad.net,
Google Code, Github, Bitbucket).  Most of them will fail 3 and 6.

3 and 2 are difficult to combine.

I get the wive that you want to host the issue tracker yourself, which
will require great effort for 2, and nontrivial effort for 5.

> To reduce the scope of this conversation, I offer the following
> anti-suggestions:
> 
> 1. no bugzilla -- it's great, I love it, doesn't meet my requirements
> for PyBlosxom
> 
> 2. no Trac -- other people like it, I don't
> 
> What options are there out there?

Roundup.  Worse than Trac, in my opinion.  Could be made to work,
perhaps, with enough effort.

I've no direct experience with other issue trackers.

> Should we go for a git-based distributed issue tracking system?
> 
> Should we roll our own system using the PyBlosxom core?

As a user I'd be happier to see efforts going into improving PyBlosxom's
core, plugins, and especially integration (a better out-of-the-box
experience with more features), than into inventing a yet another issue
tracker.  But if you find it interesting/fun, go for it!

> Are there other requirements I'm not thinking about?
> 
> I'd love to talk about this for the next couple of weeks and come to
> some kind of solution both for the software to use as well as any
> workflow suggestions that address the requirements.  Then we'll
> implement something in late October and solve this problem because it's
> seriously hampering our mojo and that sux0rz.

Why not use a hosted solution in the interim?  Pick one that lets you
export all your data easily, so you can migrate to a self-hosted tracker
later.

Personally, I like the Launchpad bug tracker.  Well, "like" is not the
right word -- it's pedestrian, doesn't invoke feelings of love, but
it gets the job done with minimum effort from my side, which I
appreciate.

Marius Gedminas
-- 
Never trust a computer you can't repair yourself.