Looking for something to hack on?

Andrew Whitworth wknight8111 at gmail.com
Tue Oct 18 18:19:14 UTC 2011


On Mon, Oct 17, 2011 at 3:14 PM, Moritz Lenz <moritz at faui2k3.org> wrote:
>> Instead, issues in the appropriate repo on Github will be created. If somebody
>> cares enough about a TT, it will get migrated to a Github issue.
>
> So we ignore the effort that people put into reporting bugs, and tell
> them "if you care, do more work for us". If you ever asked yourself why
> parrot doesn't have many users, think again.
>
> That way you only encourage potential users to write hateful blog posts
> about parrot, you don't encourage them to tell the developers what goes
> wrong.

Exactly the opposite. I think moving to use the github bug tracker
will have a positive effect on usability. This is especially true when
you consider that our trac installation is currently closed to ticket
creation from new users because of spam problems. A big portion of
this problem is that we have so far been unable to find and install a
suitable anti-spam solution for trac.

Github is a relatively popular platform, so it's likely that many
hackers already have an account there and will be able to submit
issues to Parrot seamlessly.

This is all not to mention that the maintenance burden of keeping the
trac server running is much higher than it would be on github. trac
isn't nearly as maintenance-intensive as smolder, but it's still not
perfect. OSL does a great job, all things considered, but we are
relying on their good will a lot to keep our rag-tag assortment of
software running. For a variety of reasons which I won't list here,we
shouldn't blindly think we can rely on OSL forever to host our stuff
for free.

> How about non-strawman arguments? What do you say to users who put a lot
> of effort into submitting good bug reports, and which will got lost by
> the non-migration?

I don't think there will be a non-migration, even if dukeleto is not
motivated to mass-move tickets by himself. He wants to start using the
new issue tracker. Other people have already indicated a willingness
to help with a migration, and some people have suggested automation
tools to help. If dukeleto is using the new tracker for his own todo
list and he uses it to great effect, I think that's the kind of
experiment that would be very compelling.

I personally would like to put a little bit of effort into basic
filtering so we aren't moving over a bunch of old, obsolete, junk
tickets. I also recognize that having to put human eyes on every
single one is a burden we can hardly afford.

As something of a compromise, all tickets assigned to me will be
personally inspected before any migration, whether to github. Feel
free to send any tickets my way that you think needs a human opinion
before the bots get rolling.

> How about fixing trac? Maybe by using github for authentication as an
> option?

I wasn't aware that was an option. Is that an extension which exists
already, or something that we are going to have to put developer
effort into?

Parrot does have a lot of things to do, but we can't ignore a
particular sore point (our issue tracker and server infrastructure)
just because we need to focus attention on other things (JIT, PCC,
etc). Improvements to issue tracking may actually help improve our
long-term productivity. Again, it's an experiment that I would like to
see conducted.

--Andrew Whitworth


More information about the parrot-dev mailing list