will require extra logic for source and javadocs as well. I personally like that jar info is aggregated by unclassified id