mkhitrov at gmail.com
Fri Dec 12 11:16:26 PST 2008
On Fri, Dec 12, 2008 at 1:28 PM, Glen Barber <glen.j.barber at gmail.com> wrote:
> On Fri, Dec 12, 2008 at 1:03 PM, Robert Huff <roberthuff at rcn.com> wrote:
>> "When it's ready" used to be the scheduling principle.
>> Then came 5.0 debacle: behind schedule big-time (and arguably
>> not ready when it went out the door).
>> I remember discussion afterwards, where there seemed to be
>> agreement there ought to be a more-or-less regular schedule of major
>> releases every two years (plus or minus) with minor releases every
>> few months.
>> Looking at "www.freebsd.org/releases/index.html", that's
>> getting stretched. The RC-1 announcement for 7.1, originally
>> scheduled for early September, is now listed as last week ... and
>> didn't actually happen. (Unless I missed the memo.)
> The RC-1 announcement for 7.1 did come out last week (check the
> stable@ archives).
> I personally would rather wait for quality than pushed quantity.
This discussion has come up countless number of times and the answer
is always the same - all of us would rather wait for quality, but we'd
also like some very rough timeline estimates that don't fall back into
the past. Notice that I said nothing about them having to be 100%
accurate. The questions are about the published timelines, the answers
are about the process. Hence, nothing ever gets resolved. It makes no
sense at all to have a published timeline, but claim that it is
irrelevant because "it's done when it's done." Do you not agree?
For example, RC2 builds were scheduled for 29 September 2008. When
that day comes (or same week perhaps), whoever has the ability to
change the release schedule page should update it regardless of what
happened. If RC2 builds started, that should be reflected in the
'actual' column. Otherwise, if it's a minor change in the timeline,
put the new expected date in. As is the case of 7.1 release, if the
person honestly has no idea when RC2 will happen, put in 'December',
'January', 'Second half of January'... 'Sometime next year' if it's
that uncertain. Anything at all; it takes 5 minutes to do. In the
worst case, your estimate will need to be updated again in a month or
two. In the best case, the release will be made before the expected
date. I, for one, promise not to complain about that. :)
Any date in the future will provide some information regarding the
release process, no matter how vague. Having a timeline that is in the
past provides no information whatsoever, and only irritates people who
are trying to do some planning of their own around the FreeBSD release
process. People aren't complaining because of missed dates, they are
complaining because of a lack of information; information that should
take no time or difficulty at all to provide. At least that is my
More information about the freebsd-questions