[torarnv]

Amazing summary of the mental model of the interdiff review style, and the problems with GitHub's approach to code review. Thank you!

One thing you don't touch upon (yet) is that the diff soup may lead people to prefer the squash merge strategy, to get rid of the "noise" of the fixup commits, which throws away the "good" initial 3 atomic commits as well.

With interdiff review style you're left with the initial 3 commits, and the choice of whether to land them individually or squash them is based entirely on the commits themselves and how atomic they really are.