mxnet-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Steffen Rochel <steffenroc...@gmail.com>
Subject Re: [Discussion] Separating PMC and Committership
Date Thu, 11 Oct 2018 05:20:47 GMT
Hi Haibin - thanks for driving the discussion and thanks to Isabel to
suggest review of roles and responsibilities.
I do agree with the points Sandeep raised. I though about such separation
myself and discussed with a few project members.
Honestly, I got mixed feedback and couldn't really give myself a satisfying
answer how to differentiate between committer and PMC member at this stage
of the MXNet project.

Apache MXNet (incubating) project does have an agreed process for electing
committers -
https://cwiki.apache.org/confluence/display/MXNET/Becoming+a+Committer  and
in my personal analysis there are a number of active contributors meeting
the criteria. I suggest that a fair evaluation of active contributors by
the PPMC will help to determine if established criteria are meet for
consideration to election as committer and PPMC member. Such an analysis
might also help to get a better understanding why a differentiation might
be useful and what criteria to modify or establish. It might also help to
identify potential gaps of contributions, which could be closed with
mentoring. A friend of the MXNet project suggested to look at

https://lists.apache.org/thread.html/33a6c3aa0fffa6e961aa2b861ebde333d898a5e1062d0d71d0e13d46@%3Cdev.community.apache.org%3E–
and learn from the experience of beam project in a similar situation.

Either way, it is a good discussion to have to help to grow the community
and project.
I hope the that a lot of committers and contributors will participate.

I met with Mark Thomas at ApacheCon in Montreal. He gave a pretty
interesting talk in June in Berlin -
https://foss-backstage.de/session/community-anti-patterns. He offered to
present and discuss with interested members of the project about his
experience on growing projects and communities. Happy to make arrangements,
if the community is interested.

Regards,
Steffen



On Wed, Oct 10, 2018 at 7:17 AM sandeep krishnamurthy <
sandeep.krishna98@gmail.com> wrote:

> +1
> This will be very useful for increasing our committer pool that has
> expertise in various components and not necessarily across and complete
> in-depth expertise. This will have mainly 3 major positive impact on the
> project:
> 1. Will greatly help in managing PRs and issues with fast turn around.
> There by more activity and frustration free contributions and hence more
> people eligible for committership. A positive cycle effect in community
> building. Currently this is a most critical missing piece in the project
> that is delaying and not so great quick start for new contributions.
>
> 2. Increase in committer pool is more motivation for more people to be part
> of the MXNet community and contribute and learn more parts of the project
> and increase their expertise.
>
> 3. PMC being able to focus on more critical decision making and steer the
> direction.
>
> However, like others suggested, success of this whole effort will be based
> on defining clear responsibility of PMC, committers and path for the
> community to be part of committers and PMC.
>
>
> On Wed, Oct 10, 2018, 3:37 AM Isabel Drost-Fromm <isabel@apache.org>
> wrote:
>
> >
> >
> > Am 10. Oktober 2018 11:04:30 MESZ schrieb kellen sunderland <
> > kellen.sunderland@gmail.com>:
> > > My
> > >impression is there's a high degree of knowledge and experience
> > >required to
> > >make strategic design decisions on the project.
> >
> > This statement indicates a certain understanding of what a PMC actually
> > does. As a first step, would everyone please state their perspective on
> > what their understanding of what the role committer, PMC member and PMC
> > chair actually are? I really want to make sure we are talking about the
> > same things here.
> >
> > Isabel
> >
> >
> > --
> > Diese Nachricht wurde von meinem Android-Gerät mit K-9 Mail gesendet.
> >
>

Mime
  • Unnamed multipart/alternative (inline, None, 0 bytes)
View raw message