Yet another implementation of an RDF API using Java.
In a nutshell: we store triples using SATs, which in this case are four-byte identifiers, to represent Nodes. We have two additional tables for turning a serialised form of a Node into an SAT and vice versa.
We don't attempt to do anything clever (at the moment) with environments.
There follows a list of file formats by extension, given a BerkeleyModel