Electronic spark timing ! PLEASE HELP !

[Jon]

I made an 8051 based ignition system with a pair of hall switches and 4
magnets on the V-8 flywheel.  One switch was for normal operation, the
other for starting and very slow operation.  I found that acceleration
(albeit a rather strange form of acceleration: d ^2 theta / dt d theta)
was large at slow speeds and this messed up everything.  So at slow
speeds the spark was programmed to occur immediately upon receiving a
pulse from the second Hall switch.  Here's the algorithm:  If the second
switch pulse came before the calculated time delay expired, then make a
spark and then wait for the next primary switch pulse.  I believe early
Mopars faced this same problem, and solved it in the way I described. 
My system was rock solid in timing (timing light showed no variation
whatsoever), I could put in any timing curve I wanted (even retarding at
high RPM, something weights and springs cannot do.)  I used a slightly
modified MSD 5A to power the coil.  MSD was gratious enough to tell me
how their system worked.
TVR Jon