GIT(1)
        git - the stupid content tracker

Ideas

• First Git is able to generate a hash for any piece of data. Like the coordinate of a point on a map. (Content tracking)
• Git repository hold the hash from above.
• A commit is just a piece of string that trakcs parent commit(s), author some files or trees.
• A tree is just directory which may contian files and other directories(trees).
• A new tree is created each time there is file in the tree gets updated. The new tree will link to old tree.
• Git is a mini file system using above object model to store data.
• Second a branch only records hash of certain commit, so it’s just a pointer. (Versioning)
• HEAD is a special pointer keeping track of your current branch.
• When HEAD moves, git regenerate all contents in working area by traversing trees starting from current commit.
• Third since hash is universally unique, data of that hash can be passed around network. (Distribution)

Git Worktree

It seems Git just copying current working directory somewhere else so you can have any number of working directories at the same time.

References