humble suggestion to Fedora developers

[Matthew Miller]

On Wed, Jan 23, 2013 at 09:04:55PM +0000, James Freer wrote:
> On Wed, Jan 23, 2013 at 8:29 PM, Matthew Miller <mattdm at fedoraproject.org> wrote:
> > Having some experience with timing development cycles in agile/scrum, the
> > problem with a longer release cycle is that the amount of work bitten off
> > grows to match, and you end up with the same scramble on a bigger scale,
> > actually making the problem worse rather than better.
> I wasn't suggesting a bi-annual release... but an annual one. I've

I'm not telling you what you're suggesting. I assume you know that already.
:) I'm telling you about an alternate proposal that was seriously floated
recently. That proposal involves 2-year major release cycles, with point
releases continuing at a 6-month cadence.

The annual release (or 9 month cycle, or whatever) idea still runs into the
problem I stated above: it tends towards even *more* too-much-to-do crammed
into the cycle. The fundamental problem is that estimating future work is
hard and we (humans, not Fedora in specific) are terrible at it. In fact,
we're as bad at estimating effort in hindsight as we are for the future.
Estimating over larger timeframes just makes the problem bigger. 


-- 
Matthew Miller  ☁☁☁  Fedora Cloud Architect  ☁☁☁  <mattdm at fedoraproject.org>