Sorry for extending the thread, but now I'm curious... On 11/19/15, 9:10 AM, "Todd Lipcon" <t...@cloudera.com> wrote: >I'm sure it works fine for many communities (I use CTR on some internal >infrastructure within small teams where bugs are less costly and the rate >of development is slow). But it doesn't work for all.
> >> As I said, in a CTR community there are many times where branches are >> created and the code is reviewed there before being merged because the >> authors believe the code is significant enough to require it. > > >Amusingly enough, the RTC communities I'm a part of do the opposite: you >can make a branch which operates under CTR, so long as it's reviewed >sufficiently prior to its merge into trunk. This is great for rapid >development and prototyping when a small number of contributors are >working >together on a new project. I'm curious to know, for the RTC communities, what percentage of contributions come from folks who aren't contributing as part of their day job? If I only had opportunities to contribute after work, I would gravitate to CTR communities. I keep drawing the analogy of Apache communities to these community house-building projects. I'll bet that some aspects of the house are built by professionals volunteering their time and checked by other professionals volunteering their time, but lots of pieces can be done by enthusiasts. The higher the minimum time and energy required of the enthusiast, the fewer I would expect there to be. I've been under the impression that Apache has a bias towards communities of enthusiasts, otherwise you really have a consortium. -Alex
