On Thu, Aug 09, 2012 at 04:58:55PM +0100, Laurence Hurst wrote: > On 09/08/2012 16:50, Darac Marjal wrote: > >Fair point. However, comparing[3] popcon results for xfce4-session[1] and > >gnome-session[2], less than 4% of debian users use[4] xfce, compared to a > >little over 16% who use gnome. > > > >Hmm. My numbers seem a little suspect there. Is a graphical session > >REALLY used by such a small majority of the userbase? > > > > You seem to be suggesting that xfce4-session and gnome-session are > the only desktop environments - I, for example, use pekwm at the > moment (and openbox before that) so will not be in either that 20% > (assuming no overlap, which is probably wrong). > > Looking at xserver-xorg[0], the most popular X server at the moment, > it is installed on 52% of systems partaking in the popularity > contest. I suspect this is a reasonable reflection on the number of > graphical installs in popcon as a lot of production headless servers > will probably not be reporting stats (whereas my headless boxes at > home, which do not have an X server installed, do). Because of this > I expect the real number of graphical desktop (compared to > non-desktop installs) is less than 52%. OK. Perhaps I was a little lax in my language there. xserver-xorg-core is probably the more accurate package for "uses graphics". xserver-xorg is a meta package which brings in all the drivers. But if you want a slimmer system, you'd probably remove them. You'd still use xserver-xorg-core, though. 37% of users use that regularly. A little less than twice the above numbers. So the remaining 17% (xserver-xorg-core minus gnome-session minus xfce4-session) of users use another session manager (such as kde) or no session manager at all. Still, while this is a big change, it seems like it will affect only a minority of users.
Attachment:
signature.asc
Description: Digital signature