EN
At the end of 2005 Brepols published the first constituent volume of the Dictionary of Medieval Latin from Celtic Sources (DMLCS). The present paper takes a narrative approach to explaining the lexicographical decisions that had to be made in the compilation of the lexicon. It describes how it was that DMLCS processes came to be computerized from the start (not excluding the difficulties), and the methodological differences that this made to the work. It outlines the major advantage of basing a dictionary upon a full-text database (namely the exhaustiveness it offers to the excerpting task), as well as the major challenge it poses (overwhelming the lexicographer with the sheer number of examples offered). It examines how DMLCS attempted to turn the situation to scholarly advantage in defining its objectives. The paper also describes the project's plan to combine the database with an electronic version of the dictionary as a self-interpreting compendium of Celtic Latinity. Finally, it looks at ways in which an electronic lexicon can answer questions that cannot be demanded of a conventional dictionary in book form.