FreeBSD has serious problems with focus, longevity, and
mwm at mired.org
Tue Jan 17 22:57:06 UTC 2012
On Tue, 17 Jan 2012 21:27:24 +0000
Mark Blackman <mark at exonetric.com> wrote:
> I'd have thought PC-BSD and iXsystems are the natural people to to
> take over that role in any case. The FreeBSD foundation seems less
> interested in the "for end-users" angle as well.
If that's the case, is there any reason for cutting "FreeBSD"
No, I'm serious. If FreeBSD is being run by developers for developers
(first rule of organizations: they're run for the benefit of the
people who run them), how do they benefit from a release? If users
move to some other organizations releases, and the developers don't
get any benefit from them, why do them?
On a less radical note, how about taking in the resources suggested
for the "sponsored branch", and using those to reorganize and expand
the role of release engineering? Maybe get help from PC-BSD and
iXsystems as well?
Make STABLE the "sponsored" branch owned by the expanded RE group. To
justify this, change it to an "always production ready" approach. Set
up a CI system to test it regularly, and back out changes that break
the build or tests. This does *not* include testing ports or anything
else outside the base system.
RELEASES become a snapshot of the new "always production ready" STABLE
that has ports (and anything else included that's outside the base
system) built for it and tested on it. The goal is that doing the work
to keep STABLE production ready will significantly decrease the amount
of work required to do a release.
More information about the freebsd-hackers