Open Stack

Duncan Thomas wrote:
> Agreed. It causes two problems:
> 
> 1) 9 month delays in getting code into a release

You mean 7 months ?

I find it ironic that some projects make the decision to feature-freeze
at the second milestone instead of the third and then complain about the
resulting 9 months delays.

> [...]
> Is there any evidence that our long freeze significantly increases
> stability or indeed testing? Or do people just start working on their
> features for the next release?

The RC period is the *only* time in the cycle where developers work on
critical blocking bugs, rather than their own features or their own bug
itches. So yes, it significantly increases stability. The feature-frozen
period is also the *only* moment in the cycle where integrators feel
comfortable testing, which generally results in the reporting on plenty
of bugs that somehow escaped our awesome testing automation. So yes,
that increases testing.

-- 
Thierry Carrez (ttx)