I find the whole Gnome3 debacle a spectacular failure for the gnome project.
During recent years, I can only think of OpenOffice.org as a project and name that possibly has made worse decisions and lost a greater share of users and developers. (but no, I don't have any actual numbers or surveys backing this up)
Among my friends and fellow FOSS hackers, we were almost everyone using gnome2 back in the days. Gnome3 entered and now I don't know many people who still use gnome. Users felt driven over to KDE, XFCE and other environments. I know I did, and I won't go back to gnome now on just this "promise" of a classic mode. The gnome project showed me how they didn't care much about existing users and I don't trust that they suddenly have started doing so.
Posted Nov 22, 2012 18:39 UTC (Thu) by nix (subscriber, #2304)
[Link]
But we can't add any features! It drives up the test matrix to appalling sizes and makes testing impossible!
(Actually, speak of the devil, I mentioned Emacs earlier but Rockbox is another example of a project that has huge numbers of features but somehow does not suffer a catastrophic testing collapse. This despite also having huge numbers of targets and having variable sets of features implemented on different targets. Maybe you should give the GNOME people a few lessons!)