In ordered logic programs, i.e. partially ordered sets of clauses where smaller rules carry more preference, inconsistencies, which appear as conflicts between applicable rules, are handled by satisfying more preferred rules, at the expense of defeating lesser rules. We show that this formalism can be exploited to obtain a simple implementation of abductive systems, where abducibles are assumed false by default, but weaker rules can be used to introduce them, if necessary. Moreover, the approach can be extended, without leaving the ordered programming framework, to support abductive systems involving preference, either on the set of abducibles or on the system description. The latter case appears naturally in applications such as legal reasoning where rules carry a natural precedence. However, combining preference on abducibles with a complex theory structure brings the complexity, e.g. of the relevance problem, to ¤¦¥§ , and thus such systems cannot be simulated by ordered progra...