The evolution of information processing technologies has led to a series of projects related to text corpora. In the scientific sphere, the omnipresent supremacy of English accounts for the fact that this language has been paid more attention than others in the elaboration of projects which, generally speaking, adopt a terminological standpoint. In this paper the author presents the criteria that are currently applied for the elaboration and exploitation of an electronic bilingual corpus (French-Spanish) related to a medical translation subfield (neuromuscular diseases). The availability of highly contemporary French texts allows defining an approach that is closer to the needs of professional translation from French into Spanish. The choice of a communicative standpoint, the multidisciplinary nature of the subfield selected, the creation of specific tools and their connection with a real translation project allow defining progress pathways in a space which, as explained by Habert (1999), claims a specific effort on behalf of researchers.