Is it just me or is the author just arguing for fixup commits and squashing them, something GitHub and Stash/Bitbucket handle just fine? I don't see how the idea of "publish a new version of the three commits" is new with or exclusive to Gerritt.