Comment by Retric

Programming reinvented a similar system of peer review before commits. It’s not that the reviewer is more knowledgeable about the bits of the system being changed that makes this work it’s that they have enough expertise to understand what’s involved and a new set of eyes on the problem.

In science having multiple journals acts as a safety valve here, but the underlying principle is very similar. As much as some people bad mouth it, peer review is a very low hurdle before publication that still catches a great number of mistakes.