Lennart Poettering: BSD Isn't Relevant Anymore
freebsd at edvax.de
Mon Jul 18 08:05:43 UTC 2011
On Mon, 18 Jul 2011 01:51:19 -0500, Joshua Isom wrote:
> On 7/18/2011 8:05 AM, perryh at pluto.rain.com wrote:
> > Joshua Isom<jrisom at gmail.com> wrote:
> >> On 7/17/2011 6:16 PM, Mario Lobo wrote:
> >>> On Sunday 17 July 2011 10:13:13 C. Bergstr??m wrote:
> >>>> I hope gnome does [go Linux-only].. Maybe then more
> >>>> people would forget about it and focus on making KDE better ;)
> > ...
> >> What about enlightenment?
> > For us old-timers :)
> > What's the advantage of any of these "desktop environments" (Gnome,
> > KDE, enlightenment, Xfce) over ordinary X11 with (say) FVWM2 or TWM?
> > Certainly there are some useful apps that, for better or worse, are
> > built with gtk or the KDE toolkit, but what does the full-blown
> > environment really contribute (other than bloat)?
> Desktop options are why linux has grown so well. If gnome and KDE
> didn't exist, linux wouldn't have gotten the market share it did.
Again, we see a common mixture of market share (who buys a
product or support for it), usage share (who uses a typically
free product) and mind share (who knows about typically free
In terms of market share, well... it's hard to judge about
a system that doesn't _primarily_ show up in unit sales.
> Desktop environments are a foot in the door technique for server
Although it sounds quite stupid, I have to agree. People do
want on the server what they know from the desktop. And the
way _to_ the desktop is primarily reached through GUI and
applications. Yes, it's not the OS that counts, it's the
software that allows you to get work done, and of course,
it's also the look & feel of that software.
People are different in their preference regarding the last
aspect. Some like big desktop environments like KDE, others
like things like WindowMaker. Some urgently need a desktop
full of icons, others prefer a system that stays out of their
way and lets them work. Some need good keyboard support,
others don't even touch the many complicated keys with the
This differences among users is also differences among
administrators, those who have to run the servers. Sadly,
those are often _not_ the people DECIDING about the
servers. This is mainly a task of suit-wearing (l)users
who believe in the oh holy marketing church. All the
numbers Poettering is using to "prove" his claim come
from the field of economy, of companies, of market
share. Other aspects are mostly left out.
A common problem is bloat, as it has correctly been
mentioned above. Some say that bloat isn't bloat - it's
_neccessary_ for modern application development. However,
this is highly debatable. :-)
If you see the "race conditions" in software development,
where systems get better and software gets worse, you
end up with the same "overall usage speed" (boot the
machine, start the OS, start the program, interact with
the program and so on):
hardware resources ++
overall speed = ------------------------ = const.
software requirements ++
And it's even "more const." if you are willing to agree
that those who make up the majority of "market share" are
typically users who treat their plentycore tenmelonhundred
GHz and endless disks PCs as WORSE TYPEWRITERS! :-)
In this regards, most mainstream Linusi (let alone "Windows")
could never show impressive improvements. For example, you
update FreeBSD and non-bloated applications on the _same_
hardware. What do you get? Faster "overall speed": System
comes up faster, programs run faster. Doing the same on
bloated systems, "overall speed" gets ssssslllloooowwwwweeerrrr.
In order to maintain CONSTANT speed, you need to update your
hardware. You need to do it regularly. If you don't do it,
you're out of business soon. (This is one of the aspects
that contribute to how "market share" works - this constant
renewal of otherwise fully functional parts keeps the
industry running, selling people "the same" stuff over
and over. On the other hand, it's the motor behind
development of new technology that makes today's
top technology become incredibly cheap for the masses
tomorrow, so there's no fully negative connotation here.)
And don't tell me about "advanced". There are many users
that want CERTAINITY and a constantly working environment.
They do not "advance" in the way hardware vendors, media
industry or governments want them to "advance".
> Windows clearly isn't the best server, especially older
> versions, but it's popular because desktop Windows is popular.
And the follow-up question is: _Why_ is "Windows" that popular?
Marketing and product placement strategies. Definitely NOT
quality of software.
> "server" editions of linux distributions are almost mirrors of their
> desktops, gui and all.
Yes, and I'm old enough to fail to see why I would want to
have a GUI on a server that doesn't even have a GPU. :-)
Allow me to add a very personal comment:
I'm using FreeBSD for many years now, and I have also tried
many Linusi for home use, office use, project work and even
for some critical stuff. I've always come back to FreeBSD
for most uses. This is because I'm primarily a developer.
Developers traditionally want GOOD documentation, stable
APIs and ABIs, and a system they can trust, which is willing
to give them insight to its internals. Secondary, I'm a kind
of psychologist who is able to see why _other_ systems are
so successful in many fields of IT. With some knowledge it's
not hard to conclude where development is heading.
Although there are (or have been?) some "big users" of
FreeBSD, I would say that this particular system is a
niche system. As there are many audiences in IT (to name
a few: ISPs, home commodity & entertainment users, gamers,
lamers, education, text processing offices, industrial
machine control, diagnostics & repair, mobile applications),
there _have_ to be many systems, and FreeBSD _fits_ some
niches where everything else just FAILS. There is no kind
of "one size fits all" operating system.
No system is dead that has its users. FreeBSD _has_ and
surely _will have_ users who use it, who develop it, who
help it carrying on in the future.
Poetterings generic statement "isn't relevant anymore"
can be proven wrong by _one_ counterexample (as according
to logic all allquantified statements can): FreeBSD is not
irrelevant _to me_. (And it gets even more strange, as
Poettering is allquantifying *BSD!)
And furthermore, I've found some Linux users migrating
AWAY from Linux, using FreeBSD instead. How can this be
combined with Poettering's claim?
I've really waited some time to write a statement to a
discussion that _I_ consider isnt relevant, as well as
Poettering and his creations. Please don't see this as
a persomal offence, it's _my_ individual statement as
PulseAudio, Avahi and systemd are fully irrelevant to
me. He made _his_ personal statement, I made _mine_. :-)
Happy FreeBSD user since 4.0
Andra moi ennepe, Mousa, ...
More information about the freebsd-questions