The initiative presented here was introduced as part of legal translation training (English-Spanish/Catalan) at the Universitat Jaume I (Spain) in the academic years 2001-2002 and 2002-2003. The project aimed at improving the training of future legal translators studying for their Degree in Translation and Interpreting. It aimed at fostering a certain degree of autonomy in the documentation tasks on the part of the translator-to-be by developing a textual information resource consisting of an on-line database fed with original and translated legal documents and a search engine through which the retrieval of documents was based on textual classification criteria. This paper tackles basic questions related to why this tool was needed in the classroom, why it was chosen to mix original and translated texts and why it was chosen to offer textual classification system. Moreover, the author suggests that this tool may also be used in professional translation practice by real translators in order to improve their efficiency.