What a commit history of version-controlled research papers could look like?