Abduction through Semantic Tableaux versus Abduction through Goal-Directed Proofs