Lies, damned lies and messing with popcon data
A couple of decades ago, about 60% of Linux developers used KDE, and about 60% used Gnome [Note to Windows users - switching to a different desktop environment in Linux usually requires about three mouse clicks, or you can run several at once and change which one controls the display/keyboard/mouse with <CTRL><ALT><Function Key>]. I was curious about what the current figures are, so downloaded some popcon results, which you can see below. Before that, here are some of the reasons why my results are thoroughly bogus:
1) Only machines with Debian installed report to Debian Popcon. I am sure plenty of people use Unity on Ubuntu, but only one Debian Popcon user uses Unity.
2) Popcon is optional. Lots of people do not install it, or they disable it.
3) A Debian package gets a vote if one of its executables is executed in the last month. (Some votes are missed because the executable is on a file system mounted ro or noatime.)
4) I have picked a package as a proxy for desktop environment use. I may not have selected the most appropriate package.
5) Multi-user computers get one vote per environment used, no matter how many users use each environment.
perl-base got the most votes. It is a proxy for the number of computers reporting to Debian Popcon.
xserver-xorg-core is an attempt to count desktop/laptop machines, missing out servers.
fbi is an attempt to count the number of people using Linux virtual consoles instead of Xorg.
kdelibs4c2a-trinity is Trinity (fork of KDE 3 from before the KDE developers started shitting on their users)
My attempt at getting modern figures shows 29% use Gnome, 17% use XFCE, 13% use KDE (plasma-desktop) and 43% use Xorg without Gnome or KDE. The only conclusions I trust from these figures are that both Gnome and KDE developers are good at driving their users away, but KDE developers are slightly better at it. KDE developers should not win a prize until we can find a way to compare them to TIFKAM developers.