In the present work we introduce a simple yet powerful formalism to model the evolution of knowledge bases. The semantics are based on the fundamental intuition that more recent information overrides previous knowledge, insofar as keeping the old knowledge would lead to inconsistencies. Combining this with the techniques from logic programming and ordered logic semantics yields a notion of model that nicely extends the semantics of ordinary logic programming. Our approach extends existing ones in that we model updates that contain simple facts as well as intentional information using a uniform declarative formalism that also supports what-if reasoning and transactions.