The relevance of corpora in translation studies has often been stressed in the literature during the last decade (Zanettin et al. 2003; Olohan 2004; Laviosa 2003). The advantages of corpora as complementary resources to dictionaries, terminologies, etc. have been recognised, and actually the use of corpora as translation resources and of corpus analysis software in general has become part of the syllabus of translation studies. However, the real use of corpora in translation studies still faces (some) practical problems/limitations, as already pointed out by Granger (2003): on the one hand, in some cases, sufficiently large corpora that are representative of modern language do not exist, and on the other, interfaces for accessing corpora are not user-friendly enough to satisfy the real needs of translation students and researchers. This paper deal with these kinds of problems by discussing the weak and strong points of current corpora interfaces and referring to improvements that have already been made and that should continue to be developed in the future. The chapter ends with a revision of corpus-based applications in translation training contexts and in cross-linguistic research.