Just a few days left for Rosetta, you might want to mark the date in your calendar:
On September 29, a rocket burn will essentially cancel out
Rosetta's orbital motion around the comet, initiating a
free fall from an altitude of 20 kilometers. The spacecraft
will impact the comet at a speed of about 90 centimeters
per second at 04:20 PDT / 07:20 EDT / 11:20 UTC / 13:20
CEST, give or take 20 minutes.
TL;DR: as it ventures further and further away from Earth, it won't be able to transmit much useful information, and it's unlikely the instruments will survive until the comet comes back again; letting it descend onto the surface while it's still in useful range will give a lot of up-close data.
How fast will Rosetta impact the surface?
One of the key features of the trajectory design is to
minimise the spacecraft's relative velocity at impact. The
current scenario predicts that the impact velocity will be
around 90 cm/s, around walking pace. It is worth keeping in
mind that Rosetta was not designed as a lander, and some of
its appendages including the 32m-wide solar panels will be
damaged by the impact. This energy dissipation will very
likely ensure that the escape velocity will not be exceeded
during any bounce, thus preventing Rosetta from returning
to orbit after impact with Comet 67P/C-G.
(I read from this that it is highly unlikely that the comet trajectory is affected in a noticeable way by the impact. See also the answer from symmetricsaurus above. This would qualify as a good FAQ, IMHO.)
It's fascinating how brilliant this solution is. Let it slowly bounce on the comet, so that the rebounce velocity is very low, it doesn't affect the comet trajectory in any way, and they're collecting data during the slow descent. Nice.
Rosetta couldn't really affect the comet trajectory in a meaningful way, even if we wanted to. The comet has a mass of nearly 10 trillion kg, the orbiter is <3000 kg, and P = MV. It would need to have insane velocity to impart any meaningful momentum.
No, Rosetta is already in orbit around the comet. By the principle of conservation of momentum the impact will not change the total momentum of the system (comet plus spacecraft).
The changes in trajectory happens when the engines are used on the spacecraft.
Would this still hold true if Rosetta were commanded to slam into the comet hard enough for some of the impact debris to reach escape velocity? (Not trying to nitpick your argument, it certainly holds true for the soft landing case, but I'm curious how strong this axiom is)
No. Not only because the antenna and solar panels will likely break, but Rosetta will receive a self-kill program initiated 10hrs prior to impact, which will shut off all systems on expected impact time. This is in part due to regulations regarding the deep space radio networks.
No. As far as I'm aware, there exist no regulations regarding SETI interference (mainly because this is a private endeavour and regulations would need strong international political support).
As soon as Rosetta hits the surface, its main systems will be turned off, including the attitude and control systems, as well as the main transmitter, the latter in order to meet regulations aimed at avoiding interference on deep space network communications channels.
Sorry, I have no further info on what these regulations are.
If Rosetta crashed but remained operational enough to keep an open connection but not enough to receive commands to close it, it would use up incredibly valuable resources on a highly resource-constrained network.
I don't really understand that. DSN antennas are highly directional because they have to be. If they don't aim directly at the asteroid -- and we're talking a tiny fraction of a degree of beamwidth -- they won't hear anything, even if they want to.
There is no conceivable risk of "jamming" the DSN.
I suppose it's not about "jamming" but rather "accidental overlap" with future trajectories, i.e. another mission may pass through the axis between DSN-station and comet and the signals will mix.
Be aware that the beam is conical, so even if the pointing is exact within a few fractions of a degree, the further away you are from the DSN-station, the wider is the coverage area of the beam (and the lesser the signal strength, of course).
(The rocket burn is on the 29th, the impact on the 30th, of course.)