What I'm thinking of when I say "something like Android" is that if you wanted to get millions of new people to become users (something the original article here talks about), you'd have to do something that was more green field. Few people are really interested in changing their desktop OS - what's the benefit? Nothing sufficient to offset the hassle. But lots of people were interested in trying smartphones and kindles and tablets.
Look at how much more effective iOS was as a strategy for Apple than OS X.
I agree with you that if you're setting up a computer lab for occasional walk-in use it's nice to just be more or less like Windows. Or just install Windows. GNOME 3 doesn't seem like the ideal design for that. But I don't know how the current developers think about this use-case.