Re: I thought nukes were not a solution ?
Venturing a guess - using a penetrator, instead of a proximity or surface burst. This is feasible for "leaving" rocks - for "incoming" rocks the delta-v for a penetrator to be feasible is off the scale.
If you use a penetrator, the adjacent rock is instantly vaporized providing the necessary impactor so that the energy is not wasted.
There are some tech issues with that though:
1. You need to get the speed just right. So you need a lot of more fuel to drop the delta-v down so that the nuke does not disintegrate on impact.
2. In an ideal world, you will also need a pilot probe(s) to impact the asteroid before the nuke so that you identify it correctly as rock, ice, whatever and get the delta-v down to the correct value for the penetrator reaching optimal depth
3. For a half a ton to a ton worth of nuke (assuming a 10Mt warhead) you now need several tons of armor to form the penetrator and ensure that the nuke arrives to the spec-ed depth intact. That, once again, means a boatload of fuel.
4. It definitely works only on "leaving" rocks as suggested by these guys. An incoming rock has delta-v in the 10s of km/s. We do not possess the technology to reach it in time, drop the speed to a point where the penetrator will not disintegrate and perform an attack run.