On Tue, 2012-08-28 at 12:17 -0700, Josh Andler wrote:
I'd like to proposed that after we get 0.49 out the door, we have regular planned releases. No more "always open" trunk development. But leverage merge windows instead... which also means needing to land "more complete" features in trunk than we've historically allowed. I'd love to see us have 6 month cycles to get features and fixes out to users sooner. It's awesome to have flashy new stuff for every release, but it's just not necessary.
So I like time based releases, but I'm concerned a bit about every six months. It seems to me, perhaps yearly would be better. Considering we don't developers who are paid to work on Inkscape, it might make sense to adapt the common cadence and scale it back slightly.
Then we can just version number by year ;-)
What are people's thoughts on this tried and true strategy many other projects use? Can we at least give it a shot?
I think that one issue that GNOME has had with the 6 month release cycle is that there is generally a feeling that "long term" features have a hard time in that environment. Certainly, Inkscape has it easier in that we're one codebase and GNOME features generally have to land in several libs/apps all at the same time. But, I think one thing we'd need to figure out is how to culturally support someone working on a feature like that to have a branch, get that branch tested, and be able to promote their work. Bazaar solves the technical issues, but there are some social ones as well.
--Ted