atkielski.anthony at wanadoo.fr
Fri Jan 14 13:32:11 PST 2005
Freebsd0101 at aol.com writes:
Fac> The entire point of this extended discussion, for those who have
Fac> paid attention, is that FreeBSD 4.x, which is admittedly the
Fac> fastest version available, DOES NOT work with intel's fastest CPUs
Fac> because it doesnt support the necessary chipsets ...
While I'll grant that this is an inconvenience, it doesn't seem to be
any different from any other software publisher's policies. Most
publishers will stop improving an older release family at some point in
favor of a new release family. There are both good and not-so-good
reasons for such policies. One good reason is that trying to
continually move forward with two independent release families requires
nearly twice the resources of a single family, and spreads development
resources quite thin. One not-so-good reason is that old release
families aren't as much fun to code for programmers as new families are,
and so developers like to find reasons to abandon them.
I have the same problem with other operating systems, and with other
applications. My old copy of Windows NT Server won't run on or support
many modern hardware configurations--that's what forced me to install
Windows XP on another machine. Worse yet, I can't recycle the NT machine
because some of the essential applications and hardware I use have been
abandoned in Windows XP. New versions of Windows server OSes cost far
more than the (already expensive) old versions, too. I don't see how
this is any better than the situation with FreeBSD.
Fac> ... AND, that freebsd "people" would rather ridicule people that
Fac> ask why than fix things.
People who work on FreeBSD have a rather puerile tendency to push away
anyone who says anything they don't want to hear--I'll certainly grant
that. While one can understand a certain lack of enthusiasm from a
volunteer organization (they receive nothing for their efforts, so one
can hardly expect them to jump on every problem and work three shifts to
fix it), actively rejecting anyone who doesn't say nice things is a bit
This is, IMO, the single greatest obstacles to using FreeBSD in
corporate and mission-critical environments, and it's the main reason
why I'd be extremely hesitant about recommending FreeBSD in such
environments, unless the organization in question has highly qualified
in-house technicians to support the OS. You need someone to fix the OS
urgently if a serious problem develops, and developers who get all pouty
and stop answering the phone if you don't constantly say good things
about their work are dangerously unreliable for support.
Fortunately, FreeBSD is extremely reliable. But if you are using it for
mission-critical production, you need to hire someone who can fix the OS
on the spot if something does go wrong, because you probably won't be
able to get adequate support for it from a third party.
Of course, this is true for several flavors of UNIX, not just FreeBSD.
It tends to militate against open-source software generally. Proprietary
solutions cost a fortune, but their publishers won't stomp off in a huff
just when you need them most.
Fac> So your claim that its a "heavy-duty server" platform is tainted by
Fac> the fact that in order to use the fastest server Mobos, you have to
Fac> use the slower, still-under-development 5.x. Which seems
Fac> counterproductive for an O/S that is trying to establish itself as
Fac> a choice as a server platform.
If you are using the fastest server motherboards, then you can afford to
run an operating system that is a tiny bit slower ... assuming that this
is only a temporary situation, of course. If 5.x _never_ achieves
parity with 4.x for performance, that's a much more serious problem.
New release families should always be more performant than old release
families (and don't bother to tell me that it can't be done, because I
know it _can_ be done).
One reason I've moved away from Windows is that it has consistently
bloated over its lifetime, making every new release slower than its
predecessors, and I'm not getting enough with each new release to
justify the loss of performance.
More information about the freebsd-questions