The key concern that triggered the Sentry graduation mega-thread and its
offshoots was that decisions were being made outside of mailing list by a
select few people in the hallways. That concern was put to rest by
clarification of how Sentry has been using JIRA as a primary means of
discussion. From a graduation perspective, I feel this issue is resolved.

The issue about RTC/CTR is for the project to decide and as long as code is
getting committed, contributions are welcome from all, and releases are
being made - that too is not a issue from graduation perspective.

The use of maturity model is a good thing to turn a subjective read into a
measurable/objective score and is a great idea, but as has been established
in the discussions, not a requirement for graduation.

Which leaves us to the final issue which is truly concerning - the lack of
new PPMC members. I took the time to go through the history of all PPMC
conversations on the private list to see if there was any systemic issue
stopping the inclusion of new PPMC members. What I found however was quite
to the contrary - that all the PPMC discussions have mostly been around
getting new committers in or creating board reports. The one time that the
PPMC did talk about changing a policy - it cited a thread that was already
voted on by the dev list [1]. This implies to me that the PPMC has driven
project direction by broader consensus, and thus effectively empowered all
committers and even contributors to guide it's direction already. Hence,
for Sentry moving to the model where Committer == PPMC is the right thing
for the project, and is perhaps happening implicitly already. By making
this explicit, these new committers will be able to participate in the
committer votes which will greatly benefit the project further.

Last but not least - I am aware that not all mentors of the project think
alike. Hence I don't want to make any representation on behalf of other
mentors but would strongly encourage them to participate in this thread and
share their views in case they disagree with any of my assessments above.

[1] http://s.apache.org/nyp

Regards,
Arvind Prabhakar

On Sat, Nov 21, 2015 at 3:08 AM, Lenni Kuff <[email protected]> wrote:

> Thanks for summarizing things Sravya.
>
> On Tue, Nov 17, 2015 at 11:56 AM, Sravya Tirukkovalur <[email protected]
> >
> wrote:
>
> > Hi Sentry community,
> >
> > We have had discussions about Sentry on general@ as well as dev@ around
> > its
> > readiness to graduate and current project workflows. It was good to see
> > many folks on IPMC chiming in the discussion (~150 long thread). There
> are
> > multiple discussions happening on that thread [1] , including more
> general
> > discussions about the state of the Incubator. I want to summarize some of
> > the topics discussed so we can discuss how these relate to Sentry. Would
> > also like to hear from the mentors if there is anything I missed here:
> >
> > 1. Importance of PPMC growth - should committer == PMC?
> >
>
> Agree that building a successful community involves both growing committers
> as well as growing the PPMC. We should consider how we can encourage
> community growth at both the committer and PMC level. One topic brought up
> in the general@ thread was around the distinction between Committer and
> PMC
> member. Sentry currently treats committer != PPMC. However, there are many
> other projects that treat committer == PPMC while in incubation. It's been
> two years since the project started and we have many new community members
> - it would be good to revisit whether our initial model is still what makes
> sense. Perhaps we should move to a Committer == PPMC model? This can
> probably be spun off into its own discussion thread, interested to hear
> other people's thoughts.
>
>
>
> > 2. Balance of discussions on Jira versus email
> >
>
> This was interesting. For Sentry, we have been having lots of good
> discussions on JIRA and Review Board. This seems to have worked well. One
> thing I think we should really encourage is for design discussions, roadmap
> discussions, etc all be held on the dev list rather than on JIRA. If you
> are solving a hard problem leverage the community to get input.
>
>
> > 3. CTR (Commit then Review) versus RTC (Review then commit)
> >
>
> There is a great discussion continuing on this topic on general@ right
> now.
> I'm think it is very project dependent. If you are interested I encourage
> taking a look and jumping in to this thread:
>
> http://mail-archives.apache.org/mod_mbox/incubator-general/201511.mbox/%3CCADY20s7VZzHA2BVN7oERHFA6AfeCeKj3MMtLb%2BNv-HX6uyvhkw%40mail.gmail.com%3E
>
>
>
> > 4. Should podlings fill the maturity model template.
> >
>
> IMO, this seems like a useful thing to go through prior to graduation. It
> seems like it has some good topics to think through and see how those apply
> to Sentry. I'm not in favor of having all graduating project fill this out,
> but given the healthy discussion that has been going on around Sentry
> (including some misconceptions) maybe we should do this.
>
>
> >
> > It would be good to start discussing these within our community to get
> > every ones thoughts on these topics and if we want to consider any
> changes
> > based on all the input we received. Ultimately, it's up to the community
> to
> > decide what works best for them.
> >
> > [1]:
> >
> >
> http://mail-archives.apache.org/mod_mbox/incubator-general/201511.mbox/%3C1446465555.3149570.426574697.76AAA52B%40webmail.messagingengine.com%3E
> >
> > Regards,
> > --
> > Sravya Tirukkovalur
> >
>

Reply via email to