I think it’s particularly interesting to couch one of the reasons for the Great Branch Shuffle I have heard somewhat frequently, in these terms of problem -> imagined solution -> coincidental improvment. The reason/excuse: given $currentBranchSetup, artists and designers cannot get access to a stable yet relatively current version of the game, which they need in order to work. (Programmers obviously don’t mind the game being broken, given how frequently we’re the ones to break it. Or something.)

Amusingly, I have seen this same reason espoused by advocates for movement in both directions: towards a megabranch, and towards very fine branching structures. I suppose this alone should clue people in to the fact that the branching structure itself is not the underlying issue. Either your automated testing sucks (generically: yup) or you have a cultural fear of integrations/silly legacy decisions that make integrations harder than they should be/lack of people willing to do integrations.

That’s my take on the issue anyway. I don’t actually object to the Shuffle itself (as long as they don’t occur too frequently), feeling that it’s probably interesting to see what brittle assumptions about the strategy du jour break during the move. It’s also a good time to push for changes if you have enormous binary blobs blobbing around where they need not do so, or if the aging Perforce servers are showing their age.