On Mon, Mar 23, 2015 at 5:54 AM, Florent Daigniere < nextg...@freenetproject.org> wrote:
> On Sun, 2015-03-22 at 14:11 -0500, Ian wrote: > > The way Github's code review tools are designed, you definitely want it > to > > be per branch (which will have a 1-1 relationship with a pull request), > not > > per commit. > > No disagreement here either. I'd argue that for us git and github are > the wrong tools but I completely agree with your analysis of how we're > mis-using them :) > I'm not sure, I haven't had experience with any other code review tool, but they work very well for us at my day-job, the entire team (9 experienced software engineers) seems happy with them. > You can't impose processes on people, they need to agree to them or it > > won't work. That being said, I don't know why any reasonable person > > wouldn't agree to what I've outlined. It's a tried and tested approach. > > OSS projects do; when maintainers don't like the code they just don't > merge it (and that might leads to forks and that's perfectly fine). This > is what's happening now and part of why we're in limbo. > In the early years of the project Freenet had a fairly liberal attitude to contributions, "trust but verify". I did my best to minimize red-tape for developers. This was beneficial because it meant there was a low barrier to entry for volunteers. If we didn't have this approach, I doubt many of the earliest contributors, people like Oskar Sandberg and Scott Miller, would have become involved (both started with very minor contributions). And as I mentioned previously, I just don't think we have the manpower for formalized gatekeepers (ie. we're not the Linux Kernel), so it's simply not an option to be that rigid. > What's sad is that most of the problems are from the code *paid* devs > are producing. Some might argue that it's because they're producing more > than volunteers but I don't think so. > I think we need to avoid personal criticisms of people (yes, I know that I attacked Toad earlier in this thread, but I was provoked - I should have taken the high road, I'd had a few beers at that point). > I believe that their evaluation / incentives model needs to change for > their behaviour to adapt. Maybe it's time to reconsider bounties (pay > per feature/bugfix). > I'm not sure about that. Monetary motivation works well for salespeople, I don't think it works well for engineers. It would also be a significant amount of work to administer, and I just don't think we have the manpower to do that. It could also result in a bias towards work on outwardly visible bells and whistles, and against internal stuff like refactoring. Ian. -- *Ian Clarke* / Co-Founder & CTO *OneSpot, Inc* Email: i...@onespot.com Web: http://www.onespot.com Personal Blog: http://blog.locut.us/ LinkedIn: http://www.linkedin.com/in/iancjclarke Twitter: http://twitter.com/sanity _______________________________________________ Devl mailing list Devl@freenetproject.org https://emu.freenetproject.org/cgi-bin/mailman/listinfo/devl
