I described something that seems very similar years ago[1] (in response to git-flow), however I've since come up with something better: I think knowing how to g

I described something that seems very similar years ago[1] (in response to git-flow), however I've since come up with something better: I think knowing how to get things live is important, and the entire development process should reflect that -- too many people have a "code cut" that takes days or even weeks, and that's despite using the latest and greatest CI tools. A lot of the problems may sound familiar: How do we test something? How do we know it's good enough? We have a UAT step, but it doesn't find all the problems, so do we need more tests? To be more careful? My approach[2] turns the entire problem into a software problem, and it's proving very successful to me (faster request turnaround, fewer problem reports post-deploy, and so on). Being able to select a "git branch name" for a specific user and get acceptance is powerful. [1]: https://news.ycombinator.com/item?id=6125964 https://news.ycombinator.com/item?id=6125964 [2]: https://news.ycombinator.com/item?id=11190540 https://news.ycombinator.com/item?id=11190540