The Commission ́s MT system: today and tomorrow

MT Summit VIII Santiago de Compostela (Spain), 18 – 22 September 2001 THE COMMISSION'S MT SYSTEM: TODAY AND TOMORROW Paper Type: U Angeliki Petrits, Francine Braun-Chen, Jesús Manuel Martínez García, Cameron Ross, Rosemarie Sauer, Angelo Torquati, Alain Reichling European Commission Translation Service Development of Multilingual Computer Tools 200, rue de la Loi B-1049 Brussels Belgium angelique.petrits@cec.eu.int, francine.braun-chen@cec.eu.int, cameron.ross@cec.eu.int, jesus-manuel.martinez- garcia@cec.eu.int, rosemarie.sauer-stipperger@cec.eu.int, angelo.torquati@cec.eu.int, alain.reichling@cec.eu.int Abstract This paper presents a snapshot of how the Commission's MT system (EC SYSTRAN) is used today and a glimpse of how that picture will change tomorrow. It looks in turn at: the origins of the system; how it is accessed; who requests MT and why; how users can influence the quality of output; the Rapid Post-editing Service; and the latest usage statistics, which augur well for the future. The paper closes with a look at that future, touching on the move to a new computer platform and plans for new language pairs, concluding that after twenty-five years of development, MT has become an integral part of the Commission's working environment. Keywords Machine Translation, Users, Post-editing, Future, European Commission. GENERAL BACKGROUND History The Commission has been involved in MT since 1976, when it acquired certain rights for the development and use of SYSTRAN from Peter Toma's World Translation Center in the United States. Until 1998, the Commission's version of the system – or EC SYSTRAN – was funded by the Directorate-General for the Information Society under a series of framework research programmes. However, when the system started to be used widely in- house, and therefore became an operational concern, control was passed over to the Translation Service. Development commenced with English and French as both source and target, since they are the principal working languages in the EU institutions. English-Italian and English/French-German were added in 1978 and 1982 respectively, and other language pairs were gradually included as new Member States joined the EU. Apart from including the main EU working languages, the principal development criteria were, on the one hand, the potential translation quality expected from a specific pair, and on the other, the participation of Member States. For instance, it was decided to develop French-Spanish because a satisfactory translation quality could be obtained within a short time. In contrast, English-Greek has been developed thanks to a co-financing agreement between the Greek government and the Commission. Today the Commission's MT system provides 18 language pairs (see Figure 1 below) which are accessible not only to in-house staff, but also to the other EU institutions and some public-sector bodies in the Member States. Development has always been based on internal translation needs, and EC SYSTRAN has therefore been tailored to translate Commission documents. Figure 1: Commission language pairs Access EC SYSTRAN is currently installed on a mainframe at the Commission's Data Centre in Luxembourg. Access to the system can be obtained in three different ways: 1. By e-mail. Users send their source document as an attached file to a mailbox, specifying their requirements in terms of target language(s), …

Content truncated to 500 words