pow

events come, they must have at least some work, namely 8 (every 4 work bits are equivalent to one zero in the event id).

then we check if we have room for them, we can only store up to 500.

if we have room, then fine, we take the event and store it.

if we don't have room, then we will score the incoming event against our currently stored events.

we do that score based on "timework", which is given by the work bits plus the event timestamp divided by 30 days.

this is such that old events can be replaceable by newer events over time, even if they have much more work.

and so on and so forth.