Hey Andrei,

You make some very good points.

> I do not see how the approach "let's move the project out of ASF and
> lower the bar there" benefits the _future_ of the project, compared to
> lowering the committer bar while remaining in ASF.
> (I will reconsider my vote if someone explains how the latter is less
> harmful to the project than the former)

I guess some committers could prefer this approach if they are
attached to the brand name/history of the Mesos project, and fear that
this would tarnish the legacy if lowering the bar within the ASF
causes the project quality to go down.
That's a legitimate concern, but given that the alternative risks to
effectively kill the project - especially with the legal
considerations that a fork would apparently entail - I think it's
worth trying to keep it in the ASF.
The main reason I voted in favor to moving it to the Attic was so that
interested people could actually get a chance to contribute, because
as-is, it just seems really difficult - I actually made some
contributions and want to contribute more but the absence of support
when trying to e.g. close old tickets or get MRs reviewed turned me

> To continue any development (including maintenance), the entry bar for
> the people allowed to merge code and participate in project decisions
> has to be lowered, this way or another.


FWIW, I'm a long-time committer for the reference Python
implementation (CPython), although I haven't contributed in a while.
And I remember a long time ago we had some discussions on the bar to
accept new contributors and committers, and at the time I was on the
side of being conservative and keeping the bar high. But some people
said that it was much better for the long-term viability of the
project to actually be fairly liberal and quickly grant bug tracker
permissions, and even commit permissions after a few contributions.
And they were right: it's a great way to get new people involved.
Waiting a few years to give commit rights to someone doesn't make
sense given how quickly people can change jobs, projects etc, so
having a high bar might just turn away some people would might
otherwise have become involved. And it doesn't have to affect the
quality as long as a core group of people are willing to take time to
mentor and review junior committers.
So my suggestion would be to be more liberal in granting people bug
tracker permissions and commit permissions - but of course some
minimum mentoring and code reviews would still be needed.

>  - I'm ready to provide technical mentoring (and, hopefully,
> reasonably quick reviews) in areas I'm familiar with (this,
> unfortunately, excludes most of the inner workings of the agent, for
> example)
>  - I can provide mentoring and code reviews in areas I'm less familiar
> with, but do not expect quick feedback from me there, as meaningfully
> reviewing an unfamiliar code can be 2 to 100 times slower

If you - and I assume Qian since he mentioned still being interested
and even willing to step up as chair - are willing to do some minimum
mentoring and code reviews then that would be great - I'd be happy to
submit some more MRs I want to work on, go through the tracker and
follow up/close tickets etc. Hopefully the other people who said they
were interested, like the guys from Criteo, Andreas, Javi, etc would
help too.

Also, to answer your questions, I don't really have a long-term plan
either - all I can tell you is that at work we're very happy with
Mesos: it fits exactly our needs, is rock-solid, scales perfectly well
and requires very little maintenance. It was actually the only product
fitting our criterias when we did a survey, and AFAICT this still
holds today. We have absolutely no reason to replace it.
So personally I'm happy with it continuing as a low-level distributed
operating system, doing maintenance, adding small features - or
potentially larger ones like NUMA topology support, etc. Like I said
before, I don't think curl is gaining new features every day, but I
still find it invaluable and plan to keep using it for the foreseeable



Le jeu. 8 avr. 2021 à 23:56, Andrei Sekretenko <asekrete...@gmail.com> a écrit :
> Well, one of the strengths of this project used to be effective
> collaboration between people sitting in different locations and
> different time zones (and, at certain intervals of time, working in
> different companies).
> Talking about COVID... the pandemic, assuming that it will subside,
> should probably be one more reason to vote -1 at this point in time.
> I don't know about everyone, but the additional overhead it creates
> everywhere is, for example, absolutely not helping me spend any time
> outside my working hours on any kind of knowledge transfer in Mesos
> (including discussions on the future of the project, for that matter).
> On Thu, 8 Apr 2021 at 18:29, Andreas Peters <a...@aventer.biz> wrote:
> >
> > As I see, Andrei you are sitting close to me (Hamburg, Germany), thats
> > just 30km away from me. :-) It would make mentoring more easy, if we can
> > see each other (of course, after covid).
> >
> > So, Andrei provide to be a mentor, I provide to be his student. :-)
> >
> > Cheers,
> > Andreas
> >
> >

Reply via email to