Looking on the Philosophy of kemper to provide solutions in a very creative and convenient way, I am wondering whats next in terms of "getting closer". We already have a terribly accurate speaker imprinting tech, and competition is already on the turf, so....
When looking to the profiling process, this would be chain for processing the signal:
Test signal(kemper) > Amp > Speaker > Mic > (kemper)
The kemper knows the test signal and thus can emulate the result of {Amp+Speaker+Mic} with the profiling algorythm.
With direct you get to separate the {Amp} from the {Speaker + Mic}
I know mic placement would be the last variable that kemper can not calculate, but technically, if you create a "kemper" mic with a very defined response curve (kinda difficult but not so much considering how consistent are the batches of SM57 I have had before), you technically have no variable left. Heck, you can even aproximate room response with the same mic. The algorythmn for estimating the effect from the mic would take a while to get right, but it is very possible with the right equipment (high end test heads for HiFi equipment already do this).
If done, we would have the holy grial of amp duplication: an AMP button, a CAB button, and a final MIC button, where all can be turned ON/OFF, having the most accurate repressentation ever, hearing the amp without the mic, AND maybe using MIC IMPRINTS. This for me would be the end-game for kemper.
It is a feature request but not really, as I know this might be too much hard work, but hey, if they make it one day, I would be very impressed, and very very very happy.
PD: Lets be real, pure cabinet is nice, but it doesnt cut it.