Release cycle length (was: General Resolution: Force AMD64 into Sarge)
Adam Majer <adamm@galacticasoftware.com> writes:
> Steve Langasek wrote:
>
>> <>On Thu, Jul 22, 2004 at 08:28:43PM +0200, martin f krafft wrote:
>>
>> 1 year for sarge+1 seems more likely, with a goal of 9-month cycles
>> after that. Considering how long it took the tool chain to stabilize
>> again after woody's release, I don't think it's a good idea to bet on
>> *ever* hitting 6-months for release, and I don't think everyone even
>> agrees that this is desirable.
>
>
> Short release cycles are not good. IMHO, Debian should aim for release
> cycles not shorter than one year and not too much over two years. The
> entire point of stable, AFAIK, is its code stability. When one installs
> a distribution on a server or production machine, then it is not very
> disirable to update that every few months (of course, security updates
> are a different story).
Is it better to make small scale, manageable upgrades on a production
server every 6 months, or to wait 2+ years and make an absolutely
massive upgrade in which you're jumping from 2-3 year old software to
current software?
I think it's getting nearly impossible to upgrade between Debian stable
releases due to the huge amount of changes. I would *much* rather see
stable releases every 6 months.
As it stands right now, unstable is IMO a superior distribution for
production servers (and any other type of box, really) over stable,
which is a pretty sad state...
--
You win again, gravity!
Reply to: