I suspect multiple heads on the same platter might lead to all kinds of calibration issues between the heads. You'll get weird issues like having to remember which head wrote each block of data so you know which set of calibration parameters to use to read back that data.
I could imagine certain resonances (eg. the height of the head oscillating) might differ between the two heads, effectively meaning data written with one set of resonances cannot be read back with another.
Excellent points. Also, the vibration issues with multiple heads moving independently are a bit of a nightmare, and ameliorating those nullifies some of the merely-theoretical performance gains. Multi-actuator drives have been tried before not once but multiple times, and they've never lived up to their makers' promises.
Modern drive heads are constantly reading the calibration data next to the user data and adjusting on the fly - an absolute necessity at current densities since minor temperature fluctuations would effectively destroy the ability to read or write data due to expansion/contraction pushing the head out of alignment.
IIRC the next step is to put the logic for that in the head itself with a tiny processor so the feedback cycle can be nearly instant. I suspect that would eliminate issues with one head reading something written by another.
I could imagine certain resonances (eg. the height of the head oscillating) might differ between the two heads, effectively meaning data written with one set of resonances cannot be read back with another.