The key element for both distributed storage and open source codebases appears to be refresh/review frequency.
Which is similar, but slightly different than popularity.
Essentially "With what frequency will someone put eyeballs on and attempt to use this thing"?
Below some critical threshold, integrity is compromised by freak events (hard drive crash, server goes down, host/maintainer decides to retire, etc.) and the material is corrupted/lost before it can be replicated.
E.g. re-hosting that thing that 10,000 people still have copies of, because they noticed it went down yesterday vs. that thing that disappeared a year ago and folks are just noticing
Which is similar, but slightly different than popularity.
Essentially "With what frequency will someone put eyeballs on and attempt to use this thing"?
Below some critical threshold, integrity is compromised by freak events (hard drive crash, server goes down, host/maintainer decides to retire, etc.) and the material is corrupted/lost before it can be replicated.
E.g. re-hosting that thing that 10,000 people still have copies of, because they noticed it went down yesterday vs. that thing that disappeared a year ago and folks are just noticing