Re: Architecture distribution in Debian according to popcon
Sven Luther <firstname.lastname@example.org> writes:
> On Sun, Jan 25, 2004 at 11:58:08AM +0100, Petter Reinholdtsen wrote:
> > Finally, the new popularity-contest report pages are working, and the
> > new summary scripts are available. 1684 hosts are currently
> > reporting, and this give us a good idea on which packages are in use
> > in Debian.
> > The numbers are among other things used to decide the package order
> > when building CDs, so if you want to make sure the packages you use
> > are the the first CDs, install popularity-contest. It require that
> > the host is able to send emails to participate.
> > The result pages are available <URL:http://popcon.debian.org/>.
> > New since the version in Woody, is that popularity-contest report the
> > running architecture of the participants. Currently the number of
> > samples are a bit too low to judge the real usage, but the numbers are
> > interesting anyway:
> > 1 0.09% x86_64
> > 1 0.09% m68k
> > 1 0.09% hppa
> > 1 0.09% mips
> > 3 0.26% alpha
> > 8 0.69% sparc
> > 18 1.56% powerpc
> > 1119 97.14% i386
> > 1152 100.00% total (ignored 532 without arch info)
> > Not too bad, but I suspect the fraction of non-i386 arch should be
> > larger than 2.9%. If you agree, I suggest you install and configure
> > popularity-contest to report your package usage and arch info every
> > week. Several architectures are missing from the list. Are no-one
> > using them?
> Notice that for whatever reason, the mail address which does receive the
> popcon information does believe that my machines are not using fully
> qualified domains, or something such, and thus they get rejected by
> popcon, so i somehow seriously doubt the reliability of this info. Maybe
> something to do with me using dyndns, and/or having 3 boxes behind one
> ADSL line ?
> Sven Luther
How can you tell? Does popcon complain?
I have 2 systems with popcon behind a dsl line with faked from lines
(MAILFROM=..." in /etc/popularity-contest.conf) and that seems to