Let's say impact is imminent and we need to do something about it...

It might not be a problem to send a nuke(or two) to the asteroid with conventional rocket - the type that powered Soyuz will do, or any satellite rocket should be able to deliver a sufficient punch too.

But we can't just shoot blind like that, sending a nuke onto the surface and kaboooom might cause unforeseeable results, it might send the thing right to the earth, it might have no effect, etc.

Some kind of mechanism must be implemented to scan the geographical structure of the asteroid - what is the composition, how hard/brittle is it, does it have any underground cave... etc. Which might be achieved with an optical satellite (the type that take google earth pic, i suppose). Those kind of things might take sometime to build.

And then, if we are to do something, we have to do it early, when the rock is still far far away, and still have time to change course. So we might have less than one year, or less than 6 months!

So in this 6 months somebody(like a mad scientist, a poor environmental reporter, or a rig worker) are able to put up:
- two nuke mounted rocket(one as spare)
- a sensor satellite which is able to determine the best point of impact real time, and send targeting info to the nukes
- another rocket to deliver the satellite
then we are saved.

-== or ==-

go back and sleep early, we gotta die anyway sooner or later no matter what we did.