I support giving all three active code branches a clean start, on an equal footing:
- The next release of 0.20-security (formerly expected as "0.20.205.1") to be 1.0.0, establishing branch-1.0 - The next release of 0.22 to be 2.0.0, establishing branch-2.0 - The recent release of 0.23.0 to be 3.0.0, establishing branch-3.0, from which the formerly expected "0.23.1" may be released as 3.0.1 - All three code branches to obey the established major.minor.patch versioning rules going forward. - So the next release from trunk to be 3.1.0 or 4.0.0, at the choice of the then release manager, and the pleasure of the community. Regards, --Matt On Wed, Nov 16, 2011 at 11:57 AM, Doug Cutting <cutt...@apache.org> wrote: > On 11/16/2011 10:15 AM, Scott Carey wrote: > > IMO what is important from the development and maintenance perspective is > > the _meaning_ of the > > major.minor.patch numbers as described in my previous message. > > > > If a minor version number bump means that it is a superset of the > previous > > release and is backwards compatible, then that requirement on its own > > answers whether 0.22 can become 1.1, or if it must be a 2.0 release. > > > > Whether hadoop starts using a new meaning for major.minor.patch is what > is > > of interest to me; starting at 1.x.y or 20.x.y or 999.x.y is marketing. > > Scott, this is a great point. Thanks for making it. > > > The version number is completely meaningless on its own, pure marketing. > > However, if the numbers gain meaning through a clear definition of what > > the major.minor.patch numbers signify, then there is meaning and > structure > > going forward. > > The current state of affairs seems to be: > > major: always 0 > > minor: potentially big changes; almost always breaks wire compatibility; > > occasionally breaks API backwards compatibility > > minor: typically bug fixes only; 'bug fix' not well defined; almost > never > > breaks API or wire compatibility > > Long ago I proposed such rules for Hadoop releases at: > > http://wiki.apache.org/hadoop/Roadmap > > These state that pre-1.0 releases behave roughly as above. > > > I think the community can decide two things independently: > > > > - Should 0.20.20x be renamed 1.0.y ? (perhaps not, perhaps 0.23 should > be > > 1.0 and the others left alone). > > - Should hadoop adopt a new clear definition of major.minor.patch number > > significance? > > Would you care to call a vote on one or both of these? > > > example proposal: > > * major version number increment: signifies breaks in API backwards > > compatibility and/or major architecture overhauls. > > * minor version number increment: signifies possible API changes, but > > maintains API backwards compatibility. Wire compatibility may break (see > > release notes). Included functionality is a superset of previous minor > > release. > > * patch version number increment: signifies a release where all > > improvements are fully backwards compatible with the previous patch > > version, including wire format. > > This is also similar to what the Roadmap wiki page indicates for > post-1.0 releases. > > Renaming things after the fact to try to make them consistent when the > prior rules weren't consistently followed is not easy. Instead we might > better focus on rules that we intend to obey for releases going forward > and then obey them. > > > Whatever the meaning of the numbers turns out to be will dictate whether > > releases after a 1.0.x need to be 2.0.x or can be 1.1.x > > Good point. The most accurate approach would probably be to call each > existing branch a distinct major release. Dropping the leading zero > would reduce confusion and avoid marketing but would still combine > 0.20.x and 0.20.20x which perhaps ought to be considered separate major > releases. For me this is however a reasonable tradeoff since we're > better off focusing on improving things in the future than arguing about > marketing and how to hide our past versioning mistakes. > > Doug >