On Thu, Jun 9, 2011 at 22:26, Wayne Meissner <wmeiss...@gmail.com> wrote: > On 10 June 2011 01:44, Thomas E Enebo <tom.en...@gmail.com> wrote: > >> A JRuby version 2 is like to be confusing. We may not pick 2 since it >> may confuse all the folks who think we are tracking MRI version >> numbers. Likewise 1.8 and 1.9 will probably be avoided for the same >> reason. We have been talking about this subject for years now and >> have not totally decided, but I think that is the current consensus. >> > > I think the ubuntu-style date versions aren't a bad idea. > > e.g. > If the next release is in august, it would be JRuby-11.08 or JRuby-201108 > > Version numbers are really just temporal solidification markers in a > moving stream of code, and yet people read too much into them (e.g. > the linux version 3.0 drama/non-event). I remember you and charlie > talking about the mythical 2.0 release when everything is set right in > the world, cats and dogs lie down together, and we finally throw the > ring into Mount Doom, etc for at least 4 years. > > I don't think a big-change-the-world version bump will happen - it > will be too painful for people, but I think people can deal with > occasional breakages that are communicated well better than a lot of > breakages at once.
I disagree. Ubuntu is a container of *many* packages, so it doesn't make any sense to give it a version based on features and backwards compatibility. It's the same for web applications: everyone gets the latest version, so the versions just fade away. JRuby on the other hand can control its features and backwards compatibility. It means that I can create an application/library/something and be sure that it will work on any JRuby 1.6.x. It means that JRuby can push out security and bug fixes to the 1.6-branch even after 1.7 is released. See also: http://semver.org/ // Magnus Holm --------------------------------------------------------------------- To unsubscribe from this list, please visit: http://xircles.codehaus.org/manage_email
