>>30
I give up, you're too fucking retarded to even think about how it would work.
No, honey, I can encompass every tiniest detail of each possible implementation of your idiotic proposal with my mighty intellect and conclude that it is useless at best and probably actively harmful.
As for your supposed language, well, I definitely might believe that you took some part in Perl development. Yeah, that seems totally plausible. Even more plausible than
IHBT .
>>33
Performance is not an issue, unpredictability is. It is established now that we can't check stuff at import time, therefore imports will occure at runtime. Unintended imports that doubtless will occure now and then are bad by themselves, but even intended ones will make you want to find some brick wall to bash your head against, because importing a module actually means executing its code, that, by the way, usually have side-effects
including other imports. Reasoning about cascaded and circular imports is hard by itself, now it would be truly mind-devouring.
Also, think about package-relative import paths.
Also, in fact anyone can steal a glimpse on how this retarded feature would feel: all you need is Eclipse + PyDev plugin. It has this thing about it's autocompletion, sometimes it will offer you methods from modules it knows and automatically insert `from blabla import blablabla` at the beginning of your code file. It seems like a good thing in theory, but I swear I've never seen a feature so confusing in practice - even though Pydev doesn't do it silently, like, you can actually go up and be all 'WTF IS THIS SHIT?' at what it did.