On Thu, Mar 28, 2013 at 12:11 AM, Niall Pemberton <niall.pember...@gmail.com > wrote:
> On Mon, Mar 25, 2013 at 12:12 AM, Roman Shaposhnik <r...@apache.org> wrote: > > On Sat, Mar 23, 2013 at 1:24 PM, Ted Dunning <ted.dunn...@gmail.com> > wrote: > >> One alternative to going for full-on majority voting is to recognize > that a > >> larger group is much more likely to have "noisy vetoes" by requiring > that > >> successful votes have n positive votes and m negative votes subject to > some > >> condition on n and m. Majority requires n > m, strict Apache consensus > >> requires n >= 3 and m == 0. It is easy to imagine other conditions > such as > >> n >= 4 and m <= 2 which still have some of the flavor of consensus in > that > >> a minority can block a decision, but allow forward progress even with > >> constant naysayers or occasional random vetoes. > > > > Personally, I'd suggest keeping these options in our backpocket > > and turning back to considering them in case a simple majority > > proposal runs into an opposition somehow. At this point, I'd rather > > try a simple solution first. > > I was in favour of simple majority - but a vote passing with, for > example 9+1 and 8-1 is as bad IMO as a vote failing because of alot of > +1 and only one -1. > > So I've changed my mind on this - I think it should be 3/4 majority. > This avoids a small minority stopping something, but also doesn't > completely throw out consensus. > +1 - this sounds like the most reasonable proposal of all. -- Best Regards, -- Alex