Its mostly similar to github, in that its a central place for you to talk about your data (for github its code), and there is a very strong track of each team members actions.

similarities to git; 1) we are making a portable datastore where the entirety of a body of data (metadata, discussion, data itself) can be moved around in a plain text format. 2) versioning and tracking of all evolutions to the data.

differences; obviously the diff'ing since a lot of datasets are non-text, or are very large (which we can handle).