If you understand gits internal model, it's a content-addressable data store. This happens to lend itself well to version control, but the fact that it's used for version control is somewhat incidental. It would be good if there were some way of identifying the underlying data-structures as separate from the version control system.