There's a balance between using additional screen real estate for more info and the user having to take an extra step to get that additional info, such as hovering over a button. For some people, such as myself, there is more work involved for the eyes that have to scan over that extra content everyday as compared to hovering over an unknown button once in a blue moon.
And even if they're not, I have to do mental work: icon -> wtf does this picture mean -> english translation. I find that far more tedious than standard word scanning.
