There are many differences between an ontologically inspired data model and a traditional data model, but it just occurred to me and I thought I should jot it down before it leaves me: the essential difference is in reducing complexity. By that I mean in the sense that William of Ockham meant it “entities should not be multiplied beyond necessity.” While theoretically you could arrive at a simple but complete model using object oriented, relational, UML, ERD or whatever methodology or design approach you like, in practice these approaches encourage (arouse as Tom Hite from Metalect used to say) the addition of a new attribute or a new relation or a new entity every time a new distinction is uncovered. It’s little wonder that traditional systems have such complex data structures. But it need not be so. Semantically inspired design, rigorously applied, in every case we’ve seen, dramatically reduces complexity in the delivered model.