I'm a huge critic of git, but I think it won out on merits. Git was used by the Linux kernel, and therefore from the beginning scaled better to much larger projects than other DVCS systems.
Got scales very well with source code but does terrible with binary resources like graphics. The only time I’ve seen got really fail (but still work at slowly as other systems) is when some stubborn tech lead decides to put 3rd party source deps as tarballs in git.
> Except it doesn't scale to codebases the size of Facebook and Google.
Actually, it scales well with large codebases, but the monorepo mental model pushed by FB/Google doesn't match the git workflow model. Git repo design is for a concrete entity: a single tool, binary, service, kernel, whatever. When you start bashing in multiple unrelated projects where you try to have various per-project versioning schemas, it becomes a mess.
I don't really see a benefit of having all code of an org in a single repo. Single product, sure. But whole company? Why. Not to mention, these are serious outliers.
From security standpoint, it doesn't seem great, from practical side, it's not great (bandwidth cost ect).
There was similar push by Facebook to improve Mercurial perf and they showed some impressive advantage over Git. It seems it was abandoned eventually for Git.
