Rough notes on git data model

Git data model

# Git's data model
## Snapshot
collection of files & folders within a top-level dir as a series of snapshots.
- file is a "blob", bunch of bytes
- directory is a "tree" which maps names to blobs or trees
- snapshot is a top-level tree that is being tracked

## History
- DAG
    - each snapshot refers to a "set of parents"; 
    - snapshots is called "commit" in git
    - "commits" are "immutable" in git
    - edits to commit history will create new commits & references will be updated accordingly

## Pseudocode
```
blob = array<byte>
tree = map<string, tree | blob>
commit = struct {
    parent: array<commit>
    author: string
    message: string
    snapshot: tree
}
type object = blob | tree | commit
objects = map<string, object>
```
## poking at git objects:

```
git cat-file -p SHA
git cat-file -p REF
```

## Git References
References are pointers to commits

references = map<string, string>

## Git Repository
It's just a collection of objects & references