I'm no expert on it but I like the theory behind it, because you're timing the engine based on actual behavior. A timing light tells you where the timing is. Vacuum attempts to tell you where it should be, so they're not really doing the same thing.
If you do this, I think it makes sense to warm up the engine first, and not to jumper the ALDL since you're trying to find what gives you max vacuum under real operation.
That said, I tried it once and it didn't give me a usable result. The max vacuum occurred at a point where timing was advanced enough I was getting knock. From what I read at the time this is typical. But this result does suggest it's reasonable to use the method of advancing as much as possible without knocking.
Most people say to use a timing light, and I can't really say that they're wrong, but I like the idea of calibrating against real behavior instead of a fixed spec.
Another method I've heard of is to chock the throttle partly open, then turn the distributor until you find the max RPM. That seems it would be better for performance tuning than the vacuum method.
A big caveat though. These methods are all complicated by the fact that the ECM could do all sorts of crazy things with the timing under different conditions, so what looks ideal in the driveway might not be the same while driving. And that's probably the biggest argument in favor of trusting the 25yr old EPA-approved spec.
[This message has been edited by armos (edited 04-29-2012).]