> a fork of GTK 2

For strict accuracy, MATE is not a fork of GTK2, but of GNOME 2. MATE did originally use GTK2, though without forking it, and it has since switched to GTK3 (while still keeping the GNOME 2 "look").

MATE is a fork of GNOME 2, which is a full featured desktop environment as well. MATE runs orders of magnitude faster and is much more stable than KDE. Especially on the old workstations where they installed LiMux.

I'm not going to say that the project failed entirely because of technical reasons, but at first glance it really looks like they took bad decisions. It's hard to defend a move where you end up with worse software and a worse experience for users, no matter how much money you save.

There's barely any development on the GitHub. And also it seems like there are few contributers.

Whilst KDE is a much bigger project that is actively developed.

If that's your indicative of the quality of a project then I understand your position.

First, being abandoned for new development is a great indication of the quality of a project.

Second, you haven't given us any arguments for your "indications of quality" regarding MATE and KDE.

Third, like Gnome, KDE has a huge legacy in FOSS, and is a great project in itself based on a top notch GUI backend. Some of its code even went on and become the basis of the modern web (KHTML -> Webkit -> Blink -> Node -> now also Edge), other tools like KDevelop, Krita, etc are among the best in class in what they do.

What are you, some teenage Linux nerd, with a "favorite" desktop to promote in flame wars?

And I who thought the KDE/Gnome wars was over.

MATE is not a GNOME project. When GTK3 was made and the decision to build Gnome Shell was made, the MATE project was started to fork GTK2 and the old shell.

Orders of magnitude faster? 100 times faster? Gonna need to see some numbers before I believe that.