Ten years after the publication of Discourse Pragmatics (Moeschler & Reboul), a retrospect of its principles, contributions and results seems to be necessary. In this respect, the paper highlights the ideas of the research program proposed by Moeschler and Reboul: (i) the model that seeks to capture discourse interpretation must be a pragmatic one; (ii) pragmatics must be understood as inferential pragmatics; (iii) pragmatics must integrate notions such as: global intention and anticipatory hypothesis; (iv) the identification of discourse relations (for instance, temporal reference calculus) is neither a necessary, nor a sufficient condition for discourse interpretation; (v) a pragmatic theory of discourse must be able to predict the conditions that underlie the speaker's coherence judgment.