I agree with a lot that has been said here. Primarily, that I see RTC as a way to increase committer involvement. If we know we need to be actually watching for commits to review, we are more likely to pay attention to the project and engage in discussion. I think that more involvement means a more cohesive community.
The biggest problem with RTC is that it holds up work by the most active committers, who are the life blood of the project. CTR makes it easier for the most active committers to focus on their work of moving the project forward. That part can't be overlooked. However, if there are other ways to encourage community participation, then I would propose trying those before trying to revert to RTC. On Fri, Jan 20, 2017 at 9:42 PM, Evans Ye <[email protected]> wrote: > All the other feedback seems reasonable to me. Plus I actually don't know > whether RTC can solve the problem or not. It's just a try. But I'd like to > reply in below statements specifically. > > > Cos: a mechanical +1 != positive feedback. It is as good a no feedback at > all. > > I actually don't agree with this point. > There's no mechanical +1 in our community regardless of RTC or CTR. > I think any committer in our community will at least viewed the patch and > provide a responsible +1. The only thing makes the +1 to a buggy patch is > just because of knowledge gap or human errors, which I in my opinion are > completely acceptable. > Speaking of knowledge gap, this sort of linked the story with my next > reply... > > > Roman: With all of the above -- what problem is left for us to solve with > RTC? > > One strong reason here posted by Olaf: > > > Olaf: I am miss the discussion a lot. > > With RTC all the committed stuffs have at least two person knowing that, > although there's a nondeterministic proportion of "knowing". This serves > exactly a way to sync up our knowledge. Isn't it? > > To sum up, reverting back to RTC is a proposal trying to increase > discussions in the community. > If there's a better way to do it. I'm not sticking to RTC at all. > > > 2017-01-20 7:01 GMT+08:00 Roman Shaposhnik <[email protected]>: > > > On Thu, Jan 19, 2017 at 2:22 PM, Konstantin Boudnik <[email protected]> > > wrote: > > > On Wed, Jan 18, 2017 at 01:43PM, Evans Ye wrote: > > >> > How many of us have asked for the review/feedback and when they did, > > the > > >> > feedback wasn't provided? > > >> > > >> This is more about psychology I think. > > >> For example, if a committer provided a patch and seeking for review. > > Since > > >> the committer do not highlighting anyone specifically, no one is > > >> responsible to review it. And anyhow the patch provider(committer) can > > just > > >> go ahead and commit it anyway according to CTR. So, no one feels > guilty > > >> about blocking the progress. Therefore no one will review it. > > > > > > Let me turn this around: under CTR if one is pro-actively seeking for > > feedback > > > but no one replies, the code could still be committed. Under RTC no one > > is > > > obliged to provide a feedback to a committer; but in this case the > patch > > is > > > stuck and can not be committed. And still no one fills guilty anyone, > > because > > > it is no one fault (aka tragedy of the commons). > > > > I don't really have a pony in this race, but I tend to agree with Cos > > on this one. > > > > Evans -- as you said -- it is about human psychology, but that's exactly > > why > > I don't think I have a clear understanding of why RTC provides a better > > setup. > > > > After all -- we're all committers, so there's no policy that would > prevent > > us > > from a physical act of git push'ing. Its all about self control. > > > > And from that standpoint we can have self-imposed RTC today. Suppose > > you decide that RTC is better for you -- well then just make it a habit > to > > ask for review on this list before every single commit you do. See how it > > works out for you and either keep doing it or, perhaps, start using your > > discretion of not asking when you feel really confident. > > > > This leaves disruptive pushers. Honestly, I haven't seen too much of that > > going in so perhaps I don't appreciate the full scope of the problem -- > but > > it seems to me that RTC won't solve it either. > > > > With all of the above -- what problem is left for us to solve with RTC? > > > > Thanks, > > Roman. > > >
