Too disruptive and resource intensive. People have sites to upgrade, modules to update, ...etc. If there is too much to do every 3 months, it will become a burden on everyone. Also, core developers have to put in time for making it happen, people have to test it, and Dries will go mad in no time. There is a reason Ubuntu timed it at 2 per year. Even once every 8 months is good, since it is more predictable. On 2/20/06, David Reed <dreed10@gmail.com> wrote:
My vote would be for 3 - four month cycles per year.
On 2/20/06, Khalid B <kb@2bits.com> wrote:
On 2/20/06, Dries Buytaert <dries.buytaert@gmail.com> wrote:
(Crazy idea: should 4.7 be renamed 5.0? Would it be better to call it 5.0? It would have been, but with several beta's already released it's now too late. And besides if everyone listens to Adrian R. and implements his crazy/brilliant ideas 4.7 *will* look like a point release compared to 4.8... =D
There a lot of crazy (yet cool) ideas shaping up for Drupal 4.8/5.0. People should already start preparing their patches; I hope to use much shorter development cycles in future aiming towards 2-3 releases a year. I'm thinking about trying a time-based release cycle, where development is frozen at a predefined date. It sounds like something worth evaluating. It doesn't hurt to give it a try.
A time based release cycle has merits.
It should not be more than 2 a year (one every 6 months), since it will strain the community's resources.
Ubuntu was founded because of the frustration with Debian's lengthy release cycle, and do it twice yearly.