Is this a proposal to completely reject all controversial changes?

No as there is no proposal here. See task description.

So if we decided that "yes, controversial changes should be marked as CR-2" you can't do controversial changes, which is a terrible idea. If we decided "no, controversial changes should not be marked as CR-2" then... no, we wouldn't, because that's even more of a terrible idea.

@Krenair: Gotcha - thanks a lot for explaining.
Rephrased the summary and provided an example link in the task desc.

I suspect 'controversial' is the wrong term here.

"controversial" might be wrong indeed. Quoting page 545 of "Understanding broadcast based peer review on open source software projects" by Rigby (does not seem to be freely available):

"Patches can be classified as follows:
(1) those that lead to a purely technical discussion, and
(2) those that lead to a discussion of project objectives, scope, or politics."

I meant "no consensus" in the scope of (2); (1) cannot be performed as long as (2) is not resolved.

In T115850#1793893, @Aklapper wrote:

"controversial" might be wrong indeed. Quoting page 545 of "Understanding broadcast based peer review on open source software projects" by Rigby (does not seem to be freely available):

"Patches can be classified as follows:
(1) those that lead to a purely technical discussion, and
(2) those that lead to a discussion of project objectives, scope, or politics."

I meant "no consensus" in the scope of (2); (1) cannot be performed as long as (2) is not resolved.

I can see that PDF (maybe because I'm on my university's network - we have a subscription)

Are you suggesting that we could have a policy about using CR-2 in the case of (2)?

In T115850#1794113, @Krenair wrote:

Are you suggesting that we could have a policy about using CR-2 in the case of (2)?

@Krenair: Exactly. Though I'm wondering if this might just be over-regulation.
So maybe trying to avoid such situations by improving documentation to clarify communication expectations of patch contributors in our patch contribution guidelines might be a better approach? Hmm.

In T115850#1862829, @Aklapper wrote:

In T115850#1794113, @Krenair wrote:

Are you suggesting that we could have a policy about using CR-2 in the case of (2)?

@Krenair: Exactly. Though I'm wondering if this might just be over-regulation.

Probably. What sort of policy do you have in mind?

Basically:
Should patchsets in Gerrit which first require a discussion of higher level project objectives, scope, or politics, be marked as CR-2 for the time being, with an explanatory comment asking the author to start a broader discussion about the underlying intention in an appropriate place (e.g. wikitech-l@) to clarify whether the approach taken in the patchset has broader support?

(I prefer to potentially discuss this rather unimportant item in context of T101686 hence I'm going to remove this from DevRel-December-2015 again.)

"controversial" to who?

"controversial" when it comes to project direction, whether a change is in the scope of the project, whether the trade-off of the side effects of the change is unclear.