Re: Are we losing users to Gentoo?
my answer to the subject: a few!
Dear everyone in the Debian community,
The question I want to pose today is "Are we losing users to
Gentoo?" I hate to sound like a marketing departmen drone, but I'm
becoming more and more disturbed since I'm noticing more and more
'random' outbursts on message boards about how 'cool' Gentoo
is. Whatever happened to all the Debian evangelists?
I think (even all) "Debian evangelists" are still on the side of
Debian and have not converted themselves into "Gentoo evangelists".
Whenever someone rants about Gentoo's processor optimisations
and states some overinflated performance boost such as 10%-20%, all I
can do is make a a feeble rebuttal stating that it's more like (insert
low figure without much solid evidence - e.g.. 5%) with exceptions
such as glibc, X, multimedia applications, mozilla and OpenOffice. So
then they counter that it's still an increase. Ok, so what strengths
does Debian have to make a comeback with? Unlike Gentoo, Debian has
quality assurance and security teams. We have a strict policy and bug
resolution procedures. But they won't listen and still say Gentoo.
Hmm, I've tried Gentoo 1.2 (from stage 1) cause of there's no i686
Debian tree. I've tried, nothing more - I'm still using Debian
stable for my servers and testing for my workstation. As well as
you start from stage 3 - the compile time takes longer as the
speedup saves time .. that's my opinion. I mean take X .. it takes
1 hour to compile (guess) on my Athlon 1,4 GHz. In a half year there
is a new release and I've to compile again. Does the performance
boost of 10%-20% bring in the difference between 4 minutes
installation on Debian and 1 Hour with gentoo. Imagine your system
is a PIII 500 ... and you only have X!
I know that there is a way to build an ISO on your fast machine
optimized for your (slow) target machine. But think of a security
expoit in any packege of your target machine ... you have to emerge(?)
the new fixed sources which takes time.
Yes, it's a waste of time more often than not supporting your
favourite distribution in web forums, but shouldn't Debian just be
good enough on its own that it speaks for itself? Perhaps this is what
is making Gentoo so popular of a sudden:
/me now points to Gentoo's About page prologue:
He discovered lots of up-to-date packages that could be auto-built
using the optimizations settings and build-time functionality that
he wanted, rather than what some distro creator thought would be
best for him. All of the sudden, Larry the Cow was in control. And
he liked it.
Most users changing from SuSE, RedHat or Mandrake to Debian
are happy if they can get Debian running and the system is
providing the same profit as that provided by their old
Distribution. These users are in superior number as users
which have the know-how in compiler optimizations and are
changing from Debian to Gentoo.
Silly, perhaps, but it still conveys the message that the
Gentoo user is in control. Do the cutting edge enthusiasts in Debian
have the same amount of control? Have we become so complacent at
believing that since we have the some of the strictest policies and
heaviest bug resolution/testing procedures around that we're the best
distribution around that we no longer need to seek improvements?
I'm not involved into Debian development but if it would be true
it would be sad. I assume that improvements on the policy as well as
on the technical side are taking place.
I know that there's plenty of logistical/mirroring reasons as
to why we shouldn't duplicate a lot of the i386 tree by creating a
i686 tree, but could we seriously not consider a partial i686
optimised tree as a compromise to attract some of the Gentoo users
back with our strengths in policy and testing? If not, then we need to
find something else to offer to attract the cutting-edge
enthusiast. The worst thing we could do is dismiss this
completely. Remember the days when Slackware and Yggdrasil were the
'elitist's choice'? I certainly don't ever want to see Debian even
come close to sinking.
I'm working for a mid-range ISP in germany. It would be nice to
have binary-i686 but in our case - who cares. Perhaps a shared
i686-http-Server can handle 1-n customers more but anytime we
have to put up a new machine. The advantages of saving time
with provided binary packages, secutity.debian.org (hmm, bad
example these days) stability in opposite to cutting edge
version numbers will point to Debian. I would say that if we
consider Gentoo for our system environment, the answer would
be obvious _no_.
/me throws in obligatory social contract quote to finish off:
Our Priorities are Our Users and Free Software
We will be guided by the needs of our users and the free-software
community. We will place their interests first in our priorities. We
will support the needs of our users for operation in many different
kinds of computing environment.
What if the users are companies/corporations. I don't think that
Gentoo has chances in (enterprise) companies/corporations as
PS: Apologies if my rant here is a bit disconcerted, but I've got
exams right now and typed this up fast since I was starting to loose
sleep over it = P
same for me ... and notice that I'm german ;)
The box said "Requires Win95, NT or better" and so I installed Linux.