> My code from attempt #3 was effectively using base-255, with each byte acting as a single "digit".
I think the author means base-256, not base-255.
1) Naming things 2) Cach3) Concurr invalidation ency 4) Off-by-one errors