aka j;

you can be "more sure" by using "confidence" flags.

The "confidence" flags are set by individual key routines. These routine set/reset the flags as a function of how they interpret their environment (are they getting called with bad data? or are they getting called only to find no "legitimate" stimulus?). The use of "confidence" flags allows the watchdog process to determine levels of restart (hot-warm-cold).

regards,
p