Back to 6 month schedule?

[Mike A. Harris]

Axel Thimm wrote:
> On Tue, May 23, 2006 at 12:03:30PM +0100, Paul Howarth wrote:
>> Axel Thimm wrote:
>>> I just checked with the schedule for FC6 in the wiki. I thought FC
>>> was targetting 9 months cycles, and FC6 looks like a 6 month cycle.
>>>
>>> Just curious what the targeted general schedule is, what FC6's
>>> concrete schedule is (e.g. if the general schedule is 9 month, why
>>> go 6 months for FC6?), and closely related to this, what the
>>> relationship RHEL5 to FC5/FC6 will be.
>>>
>>> My guess is that having an FC6 shortly before RHEL5 may be nice for
>>> checking some post-FC5 items that will have made it into RHEL5 (for
>>> instance xen and storage/cluster/gfs improvements). Is that the
>>> master plan?
>>>
>>> BTW in case it sounds like I would mind either way, I don't. ;)
>>>
>>> Maybe this has been discussed here before, but then I missed it
>>> when searching for "schedule" and "month" in subject lines.
>> I thought the 9 months for FC5 was always a one-off in order to get
>> the necessary installer infrastructure work done, and the plan was
>> always for 6-monthly releases in general.
> 
> I remember at the beginning of the FC5 cycle some developers that
> raised concerns against the 3 months development + 3 months bug
> fixing, and pleaded for a 6+3 model, e.g. effectively doubling the
> development efforts per cycle, while the total cycle extends for only
> 50%.
> 
> It also later went through the press that Fedora was going 9 months
> instead of 6 months, so I assumed it had been set in stone as such. :)

Fedora has always had a variable 6 month development cycle, meaning
that it is 6 months, but we may reduce or extend it if necessary to
meet certain goals for a particular release.  To the best of my
knowledge this has not changed in any way.

Fedora Core 5 was extended to 9 months under this policy, to meet
certain goals for the release.

There was indeed a lot of discussion on the mailing lists from
various Red Hat engineers, community developers and others expressing
their opinions on various hypothetical advantages and disadvantages
of making the base development cycle longer or shorter, and opinions
varied quite widely as to what the best thing to do might be.

That is no surprise really, because there is no "best" situation which
is "best" for every single person.  Making the cycle shorter, would
be "best" for some people in theory, and making it longer would make it
"best" for others out there.

While there was much discussion and opining about this, to the best of
my knowledge it was just that, and there have been no policy changes
to the length of the Fedora development cycle.  We're still on a 6
month variable cycle, which may be reduced or extended to meet
particular design/development goals for a particular release.




-- 
Mike A. Harris  *  Open Source Advocate  *  http://mharris.ca
                       Proud Canadian.