Date: Wed, 7 Jan 2004 09:57:44 -0500 From: Leo Bicknell <bicknell@ufp.org> To: freebsd-hackers@freebsd.org Subject: Re: Where is FreeBSD going? Message-ID: <20040107145744.GA74418@ussenterprise.ufp.org> In-Reply-To: <p06002027bc20fbfa66b6@[10.0.1.4]> References: <79B4EAB03B5E4649A740A8C1452F606435AF1B@y6001a.umb.corp.umb.com> <3FFAF1D4.4000709@iconoplex.co.uk> <p0600201fbc20b51ef21b@[10.0.1.4]> <3FFB4499.3050301@iconoplex.co.uk> <p06002027bc20fbfa66b6@[10.0.1.4]>
next in thread | previous in thread | raw e-mail | index | archive | help
--SLDf9lqlvOQaIe6s
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
Content-Transfer-Encoding: quoted-printable
In a message written on Wed, Jan 07, 2004 at 02:09:33AM +0100, Brad Knowles=
wrote:
> FreeBSD-5 was always going to be problematical. There have=20
> probably been more things changed for this major version than for any=20
> previous major version in history, maybe even for all previous major=20
> versions combined. They bit off a great big honking whackload with=20
> this version, and they knew it. That's why we're so far behind the=20
> original release timetable (one year? two years?).
>=20
> Any reasonable production-oriented plan would have been to stick=20
> with 4.x until such time as 5.x has been declared "STABLE", and then=20
> wait for another minor release or two after that. Timetables can=20
> (and do) slip, so you'd have to build that into the picture.
Speaking with a user hat on, I'll comment on what I believe is the
crux of the 5.x issue.
You are 100% right, in that all documentation, communication from
FreeBSD developers and soforth has pointed to remain on 4.x for
"production" machines until 5.x has a stable release, and that it
will be a while.
=46rom a practical point of view that has been rapidly breaking down
over the last 6-12 months. People need features in 5.x. Various
people have decided (for good reason, I'm not questioning the
decisions) that a large number of features go into 5.x, and because
of the difficulty in back porting don't go into 4.x. Indeed, the
only reason I'm running -current now is I need support for an Atheros
wireless card.
The take away I see is that this was too big of a chunk. The next
bite planned needs to be smaller. You can't delay one year or two
years in a production environment. New hardware needs drivers in
that time. New protocols become production deployed in that time.
I am also a firm believer that having all the developers focused so
much on meeting deadlines for all this new complexity leaves them
out of time to deal with the PR's that have been piling up.
For FreeBSD to appeal to the masses it must install on the latest
and greatest Dell or Gateway or whatever, which means it must include
drivers for today's cheaper-by-the-gross parts from China. Driver
updates in particular need to be very regular, and in the active
-STABLE release, which for now means back-ported to 4.x, even if
that means a complete rewrite because of how different the kernels
are. Otherwise people get forced to run 5.x for a few driver issues,
and then complain like crazy about all the other stuff that's not
ready for prime time.
Mom said it best, small bites, chew with your mouth closed.
--=20
Leo Bicknell - bicknell@ufp.org - CCIE 3440
PGP keys at http://www.ufp.org/~bicknell/
Read TMBG List - tmbg-list-request@tmbg.org, www.tmbg.org
--SLDf9lqlvOQaIe6s
Content-Type: application/pgp-signature
Content-Disposition: inline
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.2.1 (FreeBSD)
iD8DBQE//B5oNh6mMG5yMTYRAhc+AJ43xfQ8YoHwd82qfV1TdLc2IHWGqQCaAxBt
pP9h3KcjpEv6aBQKt33xDFc=
=Zms3
-----END PGP SIGNATURE-----
--SLDf9lqlvOQaIe6s--
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20040107145744.GA74418>