The real failure here is Mozilla, for releasing a "feature" that, as usual, appeals to computer power users, and royally fucks over computer novices.
This means that if you middle click anywhere in the window that isn't a link, it instantly redirects you to the "I'm feeling lucky" result for the contents of the last text selection you made. I cannot think of a single creepier usability clusterfuck.
Added fun: until a few years ago, http://www.microsoft.com was the first google search result for "http". This meant that malformed links (missing the colon, for instance) almost always pointed directly to Microsoft, on all platforms. It's a bit less baffling than it used to be, as not the wikipedia article on HTTP is the first result.
You should try scrolling through large source files with a mouse that has a sensitive click switch on the scroll wheel. The number of times I've found errors due to random stuff pasted into my code because I pressed too hard on the scroll wheel. grrr
Please, please if anyone knows how to disable this, let me know :/
I think ubuntu "helps" you by disabling this feature by default though.
However, what leads you to conclude they're using the address bar? The default Firefox homepage is a Google search page. You open a Firefox window and start typing and the text will be entered in the main text box on a Google search page.
I'd actually argue that Firefox's awesome bar feature is actually _for_ computer novices. Start typing 'facebook login' and before you even hunt-and-peck your way to the double O it shows a list of Facebook related pages.
Hell, what's got me confused is type in "facebook" into the Firefox address bar and press enter. It goes to http://www.facebook.com. I'm curious then - why add the "login"? (That they should be forcing ssl is a whole 'nother problem.)