One interesting thing about the Gnome Shell design process is that it has been very open and responsive. Although there was a heavy focus on creating design principles up front, the actual design has evolved a lot over the few years that the project has been active. The overview of the current design at http://live.gnome.org/GnomeShell/Design has many relevant links that you can explore if you're interested.
One thing that concerns me is that they only decided to do usability testing in Feb 2010, but according to the Wiki page Gnome 3 has been in development since Jul 2008.
Shouldn't a huge UI overhaul receive incremental user testing to stamp out bad ideas before significant dev time is invested? After an alpha is delivered, does anyone really think that if one of their big paradigm changes is found to be disliked, that they will go in and remove or significantly change it? Maybe Gnome would, I don't know--but from my experience, once code is delivered in a semi-functioning state, it's very hard to overcome that inertia to make huge directional changes. (And perhaps no big change is needed--but from these screenshots I can say that so far I personally don't like the new direction.)
I know there are plans for usability testing, but I don't know how they have progressed since this blog post http://mairin.wordpress.com/2010/02/10/gnome-shell-usability...