The case for generating URIs by hashing RDF content

In this paper we argue for using hashed URIs to represent RDF content. These URIs are generated by serializing the RDF facts which describe an Object, computing the hash of that serialization, and then using the computed hash as the Object’s URI. In this way, the hashed URI serves both as a short-hand notation for all the facts which describe it and as proof that the facts have not changed. This is particularly advantageous for distributed systems since it guarantees consistency while removing any need to trust the underlying storage. If you receive a URI, and later ask for the facts about that URI, you can prove that the returned facts exactly match those which existed at the time the URI was created.