Thanks for your long comment

1. Yes I think individual commits on a branch can get messy however the overall branch should be named as the overarching goal of what you were trying to do. The way these branches are named can also be made to follow guidelines as part of the team process. This sure beats situations with previous software when coworker X who knows everything on project Y goes on vacation and theres no documentation trail of what they've been working on or why.

2. I actually revert things quite frequently. It usually happens if I or a team mate messed up an important file. Plus with experimentation you can have each experiment version on its own branch and just delete the experiment branches if the experiments don't pan out

3. I guess you're missing my point with archiving. If master branches are used to store real-time documentation of whats in production and branches to store whats being designed - the whole company can have a searchable system of all ongoing and completed work across the entire interface