GIT - the stupid content tracker (2005)
23 points by matklad
23 points by matklad
This is just Git's readme as of the first commit. Surprisingly, but not unexpectedly, that's the best power-to-weight ratio explanation of Git I've seen, and a great study in technical writing.
Surprisingly, but not unexpectedly
As both a native english speaker and a bayesian speaker as a second language, this is an extremely 4D concept that i’m bot sure I understand.
Bounded rationality in action: of course it is expected of the first version of Git Readme to be good, because Linus is an outstanding communicator ( :) ), and because initial versions of the software put essentials into focus. But I haven't thought about that, randomly clicked on the link to the first git commit, and was surprised.
It is interesting to see comments like this one describing git:
unlike real SCM
or this one about Linux:
operating system (just a hobby, won't be big and professional like gnu)
and think about the original expectations of the authors (or author in this case) of the projects compared to where they ended up
Notably missing from this is any notion of the index, which I imagine did not exist yet. Also missing are refs and remotes. In my experience these three concepts are the biggest sources of confusion with Git, so I think it’s not a coincidence that this document is probably the least confusing, most straightforward (and as matklad said, “best power-to-weight-ratio”) explanation of Git I’ve seen.
That’s not to discredit the explanation though. It’s really good, and even though I didn’t learn anything new, I do feel a bit more clear on them, despite my experience with Git.
Notably missing from this is any notion of the index, which I imagine did not exist yet.
It’s described in the “current directory cache” section towards the end.
But yes refs and remotes are not there yet. Almost nothing is there yet!
https://github.com/git/git/tree/e83c5163316f89bfbde7d9ab23ca2e25604af290