Bill Stewart writes: > The attack also works better if you can try it multiple times with the same > numbers to work around random latency; the lowest number is closest to real. Other statistical techniques can be used to get around random network latency. Anyone who's used NTP is aware of the possibilities... .pm