Date: Tue, 16 Mar 1999 23:57:27 -0800 From: "Jordan K. Hubbard" <jkh@zippy.cdrom.com> To: Paul Apple <paula@jeffnet.org> Cc: stable@FreeBSD.ORG Subject: Re: What do people think of May 1st for a 3.2 release date? Message-ID: <12693.921657447@zippy.cdrom.com> In-Reply-To: Your message of "Tue, 16 Mar 1999 17:26:16 PST." <Pine.LNX.4.10.9903161616320.439-100000@ellie.my.domain>
next in thread | previous in thread | raw e-mail | index | archive | help
> Ideally, I'd expect two and no more than three releases per year.
Erm, we've actually been committed to a quarterly release schedule for
quite awhile now, as set down in numerous conversations both inside
and outside the project. Please keep in mind that the kinds of
releases we're talking about here are ones made from -stable, not
-current, and if we're breaking the -stable tree more than 4 times a
year then that's a process control problem that really needs to be
fixed, not a flaw in our scheduling.
I think 4 releases a year on a branch which is supposed to be dealt
with *carefully* is not at all unreasonable, with perhaps the one
"exception" being the "dot-zero" release. We always tell production
people just stay the heck away from those and we mean that, the .0
releases being aimed more at the intrepid hacker types who are willing
to deal with something that's at least a little less arbitrarily
dangerous than -current (if for no other reason other than to give
them a common reference point for general discussion and ERRATA). The
feedback that our hacker users can give us ("you broke it!") is what
we need to make the subsequent releases a little more end-user
friendly. That's actually how the process seems to work for everyone
these days - you think Microsoft is doing things any differently with
their OS products? Hell, they're making people *pay* to be their BETA
testers! Don't that beat all? :-)
We simply have to make releases or the product would never get any
better. A handful of programmers and a release engineer or two
obviously cannot test the product to the extent that many thousands of
users can, and people seem to be psychologically averse to snapshots
(if you compare the download stats) and/or need CDROM media before
they can participate in this process at all. If you want CDROM media
available then it follows that somebody's gotta make those CDs and
that somebody is going to have schedules of their own to meet if they
want to make it all happen.
To sum it up, a quarterly release schedule seems to work well, it is
achievable (with occasional quirks still to be ironed out) and perhaps
I simply need to simply codify what has always been common wisdom
among those who've been around the FreeBSD project since the early
years: If you want the very best FreeBSD on a branch, wait at least
for the .5, this being true precisely because of the stepwise
refinement afforded by the earlier releases. :)
- Jordan
To Unsubscribe: send mail to majordomo@FreeBSD.org
with "unsubscribe freebsd-stable" in the body of the message
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?12693.921657447>