Basically, it uses prefixes to identify tables and columns so that its easy to determine what tables are meant to be joined in complex queries to reconstruct objects, should one need to keep the legacy data and rewrite an app (or access it in a different way, like my current project of turning RDB data into RDF).