Version numbering for release DECISION!

[Rahul Siddharthan]

"Marc G. Fournier" <scrappy at xxxxxxx> wrote:
>yOn Sun, 27 Mar 2005, Matthew Dillon wrote:
>
>>    I've decided on a version numbering scheme.  Sorry, dates are out.
>>    As much as dates are interesting, they create problems when dealing
>>    with multiple branches.  More importantly, they do not impart the
>>    same sense of progress that real version numbers impart and, even
>>    more importantly, all kernels are moving targets and tagged with
>>    their build dates anyway so having a release date AND a build date
>>    is simply too confusing.
>>
>>    EVEN numbers denote releases.  e.g. 1.0, 1.2, 1.4, 1.6
>>    ODD numbers denote work-in-progress.  e.g. 1.1, 1.3, 1.5, 1.7
>
>Isn't that how Linux does it? *cringe*

Why *cringe*?  Anything linux does must be bad by definition?

Actually, linux at the moment seems closer to DragonFly: 2.6 has basically
become the development branch, Linus will make semi-stable releases 
2.6.x on this branch, and others will make bugfix releases 2.6.x.y for 
those who want stability.  The numbering scheme is pretty ghastly but
in effect it seems not that far from what Matt is doing now.

Rahul