incubator-general mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Reynold Xin <>
Subject Re: RTC vs CTR (was: Concerning Sentry...)
Date Mon, 23 Nov 2015 06:33:20 GMT
Most non-trivial software projects I worked on (paid or un-paid) have RTC
culture. I cannot represent every single project, but in the ones that I'm
closely involved with that use RTC, it is simply part of the culture and
recognition that mandatory code review improves code quality. (We can
debate about this in a separate thread, since this is not what this thread
is about.)

I don't think we should elevate everything to "Apache Way", "trust", or
"community building". RTC vs CTR is not about:

1. Apache Way

Given ASF doesn't require RTC vs CTR vs somewhere in between, and different
TLPs already follow different ways, I don't think any mentor or the
incubator should force their view upon incubating projects.

2. Trust

It's just part of a project's process and culture. Greg brought up that RTC
is an indication of lack of trust and committers are just treated as normal
contributors: "What I haven't seen is an explanation why a committer must
be treated the same as a drive-by. Both are subject to requiring
'permission' to make even the simplest of changes under RTC."

Committers are required to use JIRA, github, and follow many other
processes that "drive-by" should follow. I don't see why "code review" is
different from filing JIRA tickets. In most RTC projects, committers do
have more rights -- a committer can review somebody else's patch and commit

3. community building

Lots of successful open source projects, both inside and outside ASF,
employ RTC. As Todd mentioned, almost all the top 10 most starred (on
github) projects use some form of RTC, so it is hard for me to believe that
RTC would hinder community building. Of course, one can always argue that
if those projects had employed CTR, maybe they would've been even more
popular. But then we got into the area that we just have to agree to

On Sun, Nov 22, 2015 at 9:37 PM, Todd Lipcon <> wrote:

> On Sun, Nov 22, 2015 at 12:18 PM, Konstantin Boudnik <>
> wrote:
> >  >
> > > > The question is not to decide if C-T-R is The Apache Way over R-T-C.
> > The
> > > > question is wether a project entering incubation with a selected
> R-T-C
> > > > mode is likely to exit incubation for the simple reason it will be
> very
> > > > hard for this project to grow its community due to this choice. It's
> > > > like starting a 100m race with a 20kb backpack on your shoulder...
> > > >
> > >
> > > If you have any statistics that show this to be the case, I'd be very
> > > interested. RTC is the norm in basically every Apache project I've
> been a
> > > part of, many of which have thriving communities and are generally
> > regarded
> > > as successful software projects.
> >
> > Do you have any statistics on that, Todd? Would be very interesting to
> see,
> > indeed.
> >
> >
> I don't have incubator stats... nor do I have a good way to measure "most
> active" or "most successful" projects in the ASF (seems that itself could
> be a 'centithread'-worthy discussion). But a potential proxy could be the
> number of stars on github:
>  (sort by number of stars)
> Of the top ten:
> Spark: RTC via github pull request
> Storm: RTC ( see "Code
> Change")
> Cassandra: RTC (based on my skimming the commit log which has "Reviewed by"
> quite often)
> CouchDB: RTC ( see "RTC" section)
> Kafka: RTC (based on "Reviewed by" showing up in recent commit logs)
> Thrift: CTR
> Mesos: RTC (based on reviewboard links in most of the recent commits)
> Zookeeper: RTC (based on personal experience and comments above in this
> thread)
> Cordova: CTR (based on
> )
> Hadoop: RTC (based on personal experience)
> Briefly looking through the #11 through #30 projects I also see a
> substantial number which operate on RTC (and others for which I don't know)
> So, I don't think there's much evidence that RTC prevents a project from
> becoming successful in the eyes of the developer community. Also worth
> noting that several of these are relatively new TLPs (i.e. within the last
> ~3 years) whereas others are quite old but still active and successful.
> -Todd

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