Summary from yesterdays (mini) FESCo meeting

[Michael Schwendt]

On Sun, 31 Dec 2006 05:26:02 -0600, Callum Lerwick wrote:

> > They learn best from submitting packages, from applying the guidelines to
> > their own packages, and from trying to review packages. This is what they
> > do once they maintain their package in CVS on their own without any peer
> > reviewing. They cannot learn from a brief list of YES/NO answers or
> > MUST/SHOULD items. Example:
> 
> The problem with this, is in reality this can not scale. Not everyone
> can submit and maintain packages.

How does that contradict with what I've written above? I talk about how
the possibly silent observers of bugzilla traffic can learn from stuff
that is talked about during reviews, while you talk about scalability.
Packaging needs hands-on experience. Reviewing needs hands-on experience.
You cannot learn techniques from skimming over checklists. You need
documentation (or guidance) on how to perform the checks yourself plus the
creativity to check things beyond predefined checklists.

> And people busy maintaining packages
> don't tend to be interested or have the time to do reviews. Hell, I'll
> admit that personally, I signed up to package stuff. Not to do reviews.
> (I think I squeezed in just before the "must review other people's
> packages" policy solidified

I don't think it's a policy. It's supposed to encourage contributors to
exchange reviews and at the same time demonstrate some of their
skills. But I've never backed it up, because I don't think it works
well. We have "sponsors". We don't need to force new contributors into
looking for things they might review before they are permitted to maintain
their packages in FE. You can require a new contributors to perform a few
reviews reluctantly, then sponsor him, but some months later the
contributor orphans his packages because of lack of interest or time,
anyway. So, when put up the extra hurdle in front of the new contributors?
Sponsor people as soon as they manage to push a new package through the
review process, and assist them in case problems turn up.

> and makes me feel a bit like I got in easy,
> but hopefully I've made up for that by now.) Doing reviews is of little
> interest to me, I've reviewed only a handfull of stuff I really
> wanted/needed and that no one else was in any rush to review.

For me it's different. I don't "own" more than a hand full of packages,
because in my point of view it's better if we avoided the
50-packages-per-owner scheme and built many more teams that share the
workload. However, interest in doing reviews is lowered due to metrics and
systems that aim at quantity instead of quality.