> -----Original Message-----
> From: email@example.com
> [mailto:firstname.lastname@example.org]On Behalf Of Steven
> Sent: Monday, May 30, 2005 7:26 AM
> To: Ted Mittelstaedt
> Cc: email@example.com
> Subject: Re: 4.11-RELEASE install error
> I don't want to offend anyone, but is this the official
> FreeBSD.org position?
> I had hoped that 4.11 would be the crown jewel of the 4.x
> series and would be
> worth buying on CD or DVD. I know that what Ted says is the
> truth. I've went thru various installs, selecting different
> configs each time
> and I noticed that KDE 3.4 wasn't in 4.11-RELEASE.
I think the entire FIrefox thing was a big embarassment. The ports
tree that went on the 4.11 ISO called for Firefox 1.0 if I recall
and literally within weeks of the release the Firefox team found
a giant security hole, and knee-jerk pulled all 1.0 code off
their web and FTP sites. I think that since then they have put
it back, to satisfy to large number of people still pulling the
code down via automated means. (like FreeBSD ports)
Then to compound the problem when the Firefox team released FF 1.01,
1.02 and such, they decided to make some function calls that were not
in the version of freetype that the 4.11 ISO had as a precompiled
binary. So the upshot is that if you installed X Windows from the
precompiled binary it used the old version of freetype, and when
you tried building the newer patched version of firefox, it blew
chunks with a bunch of link errors about missing functions.
Both of these goofs were the Firefox team's problem, not the
FreeBSD teams problem. Unfortunately they were a very visible
marring of the 4.11 ISO since Firefox is one of the few web
browsers that is available for UNIX that won't fold up and die
when it encounters various Microsoft constructs in people's
websites. While it may be satisfying to tell people to just not
surf www.cnn.com or some such with their old versions of Netscape
because the HTML is wrong, it's not very useful.
> I wish the core team saw each release as a milestone captured
> in posterity.
> Anyway, I performed a couple installs last night directly from
> CD rather than
> the ftp directory that I created with cp -R /cdrom/* .
> I notice that during a CD install, the progress bar sometimes
> appears to back
> up, perhaps indicating a read that has to be performed again.
> Perhaps when I
> use cp, there isn't anything detecting these read errors and I
> get corrupt
> So to check out this theory, I downloaded the ISOs again and
> I'm going to
> mount them and cp into the directory again.
> I love Unix in general and FreeBSD in particular. I wish the
> core team
> wouldn't turn their back on the 4.x series so quickly.
> Please, please,
> please schedule a 4.12 release that's as pristine as possible.
It is really a matter of allocation of resources I am afraid. The
problem is that 4.x lacks some extremely important things, like
threading, that are absolutely necessary to get good performance out
of some of the most visible Open Source applications. In particular
I'm thinking of mysql, which runs like crap in a non-threading
You take mysql and stack it up against postgres on 2 identical
4.x SMP systems, and postgres will kick the stuffing out of mysql.
Unfortunately the morons that do various bakeoffs for the national
media never seem to use postgres when they are writing their Linux vs
FreeBSD articles. Instead they use mysql which has been optimized
for threading, and Linux has a much better threading package than
For this and other reasons the team wisely decided to do an entirely
separate FreeBSD version train, rather than try to stuff all these
fundamental kernel changes into 4.X
The problem that the Project has run into, though, is that hardware
has changed ferociously in the last couple years. Each time another
4.X and 5.X release came out, more and more newer systems would run
better under 5.X, and more and more older systems would run better
under 4.X. In the meantime pricing is plummeting into the toilet,
to where today you can walk into a computer store and have a server
custom-built with IDE raid for under $1000, that will kick the literal
crap out of a 4 year old server bought from Dell for $15,000. So
it is quite obvious that it is imperative to get FreeBSD running as
good as possible on this plethora of wonderful powerful NEW hardware
that is coming out daily, so that when the national ragazine writers
evaluate FreeBSD, that FreeBSD comes off with high marks. Thus,
we need to concentrate all our development time on the newer release
and let the older release end. The 4.X train really has been carried
far longer than any previous release train of FreeBSD.
4.X runs fantastically well on that elderly server you have that was
manufactured in 1999, 2000, or 2001, and paid $15,000 for. But
you can't let the emotional pull of having spent that $15,000 four
years ago on a box, hold you back from replacing it with a brand
new system that is more powerful. There are too many other people
out there who are buying those new boxes, and they all want FreeBSD
to run well on their stuff, same as you do. And the unfortunate facts
of the industry is that because they just spent their $1000 last week,
what they want matters more than what you want, who spent your $15,000
4 years ago.
The computer industry is still very immature and very much has the
mentality of I don't care what you did for me a year ago, what did
you do for me yesterday, and what are you going to do for me tomorrow?
It is not like the automobile or appliance industry where they are
used to far longer customer support times. Hell, you can still to this
day go out and drive into Ford and get a brand new factory-manufactured
302 engine dropped into your 1966 Ford Fairlane, and that is a 39 year
old car. Imagine trying to get repair parts for a 39 year old computer