Microcontroller decoupling, >1uF rules?

The problem with that thinking is that caps are very seldom connected to the chip by traces, rather by ground and power planes. I recall all manner of analysis that talked about the geometry of the three way connection of chip, cap and planes, mostly wrong analysis because it was never tested or even simulated other than in the mind.

That has been debunked. The caps are coupled to the power plane and provide bulk capacitance more so than being individual caps to the pins. A cap can be inches away before you see significant effects. This is because the power and ground planes act as transmission lines supplying the current required until the current wavefront reaches the cap and the cap provides current in the reflection. Transmission lines are not lumped inductors or capacitors and thinking of them as such produces wrong results. Most rules of thumb have never been tested and when they are tested turn out to be wrong or at least exaggerated.

Do you analyzed transmission lines by treating them as lumped inductors??? Why would you treat the transmission line formed by the power/ground plane pair as a lumped inductor between the power pin on the chip and a decoupling cap? The power/ground plane pair is not just another cap connected to your PDS, it's the transmission line connecting everything together.