Thilo Pfennig wrote:
Rahul Sundaram schrieb:
> Impressions based on arbitrary numbers and random hunches are not very
> useful. We are looking at collecting more concrete stats instead.
Well absolute numbers are not very useful, especially when counted
worldwide. You get a better sense of a distribution if you know what
Counting users, collecting hardware information etc is *very useful*
exercise in itself that is different from surveys etc. Basic information
that we collect currently is purely anonymous server side
infrastructure. Analysis of these metrics can almost be completely
We have also thoought about those problems at GNOME
marketing list http://live.gnome.org/CountingUsers
The problems is
quality of the data.
Yes, the metrics page discusses the pros and cons of different
approaches in detail.
I think it would make much more sense if this would
be a combined effort that would be led by freedesktop.org
anyway we need measures that mist distributions or desktops accept. If
all do it differently you wil not have any relations you can count on.
So I would rather recommend to discuss such things on freedesktop-promo
Our current efforts on collecting metrics is not limited to the desktop
and is not just a marketing effort. The particular list seems to be all
about coordination between promoting desktop environments like GNOME,
KDE etc which is important but doesnt seem very related to what we are
trying to do here.