Re: [gentoo-project] Questions to nominees

[Michał Górny]

On Sun, 2021-06-20 at 23:40 +0200, David Seifert wrote:
> Nominees,
> congratulations on your nominations! As part of this year's elections,
> I'd like to pose five questions to the nominees, that I believe are
> important factors in considering someone a good candidate for the
> council:
>
> 1. Do you feel you have enough time to commit to serving as a Gentoo
> council member in the 2021/2022 term? Does your commit activity support
> this?

Yes. I consider this very important, and I wouldn't have accepted my
nomination if I couldn't have found sufficient time for Gentoo. You need
to stay on top of things, and others needs to be able to talk to you.

> If you served in 2020/2021, have you prepared for council meetings
> and finished all unfinished business for which you were responsible (as
> a council member)?

I've served in 2017/18, and I didn't have unfinished business then.

> 2. Project X and Project Y have irreconcilable differences, but you
> aren't involved with any of the projects. A crucial technical decision
> needs to be made. How will you react?

If the projects can't reach a consensus themselves, the Council should
work with them in order to understand the problem and establish
a working solution. It is not always possible to find a solution that
works with everyone but the Council shouldn't just arbitrarily decide
for the projects without proper research.

It's important not only to consider the perceivable advantages
and disadvantages of any of the options but also the background,
workflows, implementation and maintenance costs, community response.
A seemingly good decision is meaningless if it's not going to be
implemented (see e.g. how QA team arbitrarily tried to enforce Qt-style
versioning on GTK+ flags in the past).

> Will you defer?

Ideally, the Council should be able to analyze the situation timely
and make a decision. However, in reality I can imagine that sometimes
deferring is the most reasonable option (e.g. when a new interesting
point has been made in the discussion).

> Do you consider
> abstaining a viable option for the group of people making decisions as a
> last resort?

By accepting a Council position, you accept the responsibility coming
with it. Abstaining for a last resort decision means neglecting this
responsibility. So, no.

> 3. Given your typical area of responsibility, how have you performed?

I think that this point my most typical area of expertise is Python.
I believe we've made a lot of important progress over the years.
The switch to Python 3.9 went smoothly (though I admit things weren't
perfect). Python 3.10 isn't going to reach RCs until September but we've
actually managed to enable its support on a large subset of packages. We
no longer have major stabilization delays, test coverage has improved
greatly, a large number of packages is being bumped timely. We have
really good documentation. A lot of new contributors are helping out.

> 4. What positive change/idea/plan do you have for Gentoo that you would
> try to further (not necessarily as a council member)? By positive change
> I mean actually changing something concrete, not some diffuse notion of
> "improving how the council acts" or non-tangible deliverable.

I think the key point is to focus on our strengths, and reiterate what
strengths we have today rather than clinging to the past.

I believe that Python support in Gentoo is superior to any other
distribution. It's important not to lose that edge, and push forward.
This means up-to-date packages, high test coverage, timely porting to
new implementations.

Getting new contributors is very important. However, I don't believe we
should be pursuing the 'everyone must become a developer' model anymore.
GURU's been a great example how helpful it can be to just let people do
their thing. People are working together, learning and becoming good dev
candidates. It somewhat resembles the Sunrise of old, except that it
comes with practically no maintenance effort from developers, and that
makes it much more sustainable.

I also support the idea of better defaults. While choice is often good,
forcing people to make choices all the time is not. If I develop
in Python, I enjoy being able to control implementations precisely;
if I just want to install a random program, ideally Gentoo should let me
do that without bothering me about Python versions. We should aim for
flexibility but also for reasonable defaults.

This is also a prerequisite for reasonably good binary package support.
With Gentoo's packaging model we can't expect to compete with binary
distros but we can make lives of some of our users much easier. Both
in the terms of providing binary packages ourselves and in making it
easier to deploy local binary packages on site.

> 5. Do you think the council should be more agile - i.e. take decisions
> for the purpose of propelling Gentoo forward, rather than waiting for
> the decision to be made for it?

Individuals propel Gentoo forward, and the Council should do their best
to support them. Of course that doesn't stop Council members from being
these individuals, I just don't think we should consider Council members
special in this regard.

> Would you consider a small number of
> departing views on the mailing list or IRC to be enough to derail a
> proposal?

Every decision breaks somebody's workflow. We need to accept that
sometimes some people would be unhappy with what we do (or in extreme
cases even threaten to leave Gentoo). However, sometimes their arguments
can actually be important and are worth considering.

> When do you consider a controversial issue to have been
> discussed enough?

At the point that the same arguments start being reiterated over
and over again, and nobody's making new points anymore.

-- 
Best regards,
Michał Górny