A Generalization of Version Control
- A version control system is a system that keeps track of changes to one thing: Data.
- It essence, it can be represented with one structure:
Change. This type represents the new version of the data, the old version of the data, and how the data changed.
set Change { Data : Type } {
previous: Option (Change Data),
transform: Data -> Data,
result: Data,
}
This more concrete representation might look something like this.
struct Change<Data, Transform> {
previous: Option<Hash<Connection<Data, Transform>>>,
transform: Hash<Transform>,
result: Hash<Data>,
}How would this generalize something like Git?
Git is just a bunch of files, directories (trees) all compressed and stored as content-addressed blobs in the objects directory. It can also further save space by packing different versions of files together using a binary delta.
// Encodes change to object. i.e. a file or directory list
struct ObjectChange {
previous: Option<ObjectChangeHash>,
transform: HashOfDiffAlgorithm
result: ObjectHash
}
struct CommitChange {
previous: Option<CommitChangeHash>,
transform: (),
result: CommitHash
}- Pull requests and issues can be opened on any file or directory.
- Since everything is a content-addressed DAG, git repositories can contain other git repositories and share data with other git repositories naturally.
- People will be able to see using reverse-hash-lookup newer versions of shared files and be able to update them in their own repository.