It's true: Try as I have, I just can't learn to love Unity. The concept itself is fine and I can even live with its biting lack of customizability, but I just can't take the bugginess anymore. Random things happen when I try to switch between applications -- trying to open recently used files via the dash launches Nautilus instead, and the whole thing just generally doesn't work the way I need it to on a production machine.
True, these may be bugs related to my particular configuration. Also, I should be a good open source citizen by filing reports on Launchpad and all that. But I tend to neglect my civic duties, and I'm tired of blaming myself for the shortcomings of software that I didn't engineer. I just need stuff to work -- like it did in GNOME 2 -- and Unity doesn't.
Since I've fared little better with GNOME Shell, even after attempting to use it on both Ubuntu and Fedora, and neither Xfce nor KDE appeals to me very much, I figured I'd see what MATE could do by installing it on my Ubuntu 11.10 system.
Running MATEThere are some PPAs available for MATE, but because the one for Ubuntu is not complete and my CPU needed a workout, I compiled it from the GIT source by following instructions on the Ubuntu forums. They were pretty straightforward, even for a decidedly non-elite hacker such as myself, and all but a couple of non-essential components built successfully.
After compiling and installing, I selected MATE in Ubuntu's lightdm login screen and was relieved to be presented with a desktop environment that actually conforms to the conventions that have dominated computer interfaces for two decades:
I won't bother uploading any more screenshots because as the ones above make clear, MATE looks pretty identical to GNOME 2. Besides a few very minor interface changes and a different theme, you'd never know this wasn't just Ubuntu back in the glory days, before Unity came along and shattered my dreams.
I didn't expect to see anything very new in MATE, though, since the revision history of the code indicates few changes -- and all of them are the work of the project's founder, perberos. Obviously MATE does not have the enormous development resources of the GNOME project behind it like its predecessor did.
But even if MATE diverges little in look and feel from GNOME 2, that's not a bad thing. After all, as I've said, GNOME 2 just worked, which -- surprising as it may sound to the folks at Canonical and GNOME -- is what I look for most in a desktop environment these days. GNOME 2 may not have been ideal for touchscreens or tablets, and it wasn't the most visually dazzling interface out there. But it got the job done without giving me a headache or turning every mouse click into a surprise by eliciting totally unpredictable behavior.
Clearly, MATE will need some help from other developers if the project is to become truly viable. It's doubtful that the lone hacker currently working on the code will be able to keep up with security and compatibility issues, let alone adding new features. Endorsement from mainstream Linux distributions also wouldn't hurt MATE's prospects, but that seems like a long way off given that the project doesn't even have a real PPA yet.
Nonetheless, the interest in MATE seems to be there among users disgruntled with Unity and GNOME 3. If some of them are able to offer the skills and time necessary to turn MATE into a serious contender in the world of open source user interfaces, GNOME 2 may turn out to be not as dead and gone as Canonical and the GNOME people would have us believe.
N.B.: for what they're worth, I copied the Debian packages that I produced while compiling Mate to my web server. They're for amd64 architectures only and I make no guarantees that they'll work for you, but anyone interested can download them here.