Great starting points!

I agree w/ JB's point that some of this should be private@ discussion. I am
keeping my comments to things that I think everyone on dev@ should be part
of.

For all mentions of ASF policy: we should have our own link to the relevant
policy. It is hard enough for us to find ASF policy docs, and almost
impossible to know whether docs exist, whether you have the "right" ones or
"all" of them.

On Tue, Jan 23, 2018 at 8:43 AM, Ismaël Mejía <[email protected]> wrote:

>
> * Code of conduct
>

+1 to link to ASF policy

* Proposal process
>

+1 to link to ASF policy "it has to happen on list"

Design docs fall into the same category and coding and code review: use the
right tool. It would help to describe our practices and tools. I want to
see comments like: "let's move this from a PR to dev@" and "this needs a
design doc" and "these doc comments should move back to dev@" and also
"this is small enough you can just code it up".


> * Governance model
>

I think the biggest gaps we have are in governance of governance. ASF
policies establish about half of what is necessary, notably missing the
need for expiry and renewal. Mostly, we should be transparent about what
the PMC does and how often, etc. Threads on private@ of course are private
for a reason, but they should be private without seeming arbitrary or
mysterious.

* Contribution policies



- Avoid CI issues: Extremely long tests that break for unrelated
> issues should be minimized
>

+1 and this is in crisis IMO. Since "hope is not a strategy" I think we
need to cut scope and rebuild in a way where signals don't clobber each
other.


> - Review of pull requests should take less time than the current
> average (We should probably get some stats on this / define what is an
> acceptable review time and what is the expected process when a
> reviewer is not active).
>

+0.5 I think we should reduce latency for response and document policy so
newcomers feel empowered to "ping".

-1 to any lower review standards. Discussion on a PR should build
community. You can also issue a PR to a PR and discuss *that* PR to show
even more outreach. There is no conflict between high quality code and
friendly community. Anyhow for me, I love to write, read, and discuss code
:-)

- We should add the previously discussed policy on stale PRs to the
> contribution guide.
>

+1

- Encourage a well balanced distribution of reviews, there are some
> committers that do many reviews, and of course this is not ideal.
>

+1 can we pull numbers or dashboard review load over time?

- Give priority for reviews / help to new contributors that are
> starting in the project or to critical fixes.
>

+1 especially new contributors

- ... commitership/PMC status...
> https://flink.apache.org/how-to-contribute.html#how-to-become-a-committer


+1 to link to ASF policy, maybe add something, maybe not.

Kenn

Reply via email to