The innovative pilot program to establish a Translation Technology Service Centre for the Greek public sector was launched in 1994. The Centre was connected to EC-SYSTRAN via X25 and staffed by two linguists and an engineer, all of whom were trained in Luxembourg. One of the linguists was also specialized in public relations so as to promote the system. This pilot programme can be seen as something special as it was the very first time the European Commission had made ECSYSTRAN available to the public sector of one of the Member States. At the same time, it was especially innovative for Greece, as the technological infrastructure of the public sector was quite poor. The first year was the most difficult, and the hardest problem encountered was to explain the notion of machine translation, its purpose, its limits and capabilities.
This proved to be harder than expected since information technology is not widespread in the Greek public sector. Even the computer-literate had a lot of reservations as machine translation is not widely known and often misunderstood. Having users that are completely unaware of what new technologies offer can cause many problems, since they do not understand the procedure and the importance of an “error-free” input text. Through lack of awareness, it was not uncommon for them to hand in hard copies even when they had the text in electronic form. But if this was to their disadvantage in terms of slowing down the translation process, things became even worse when it came to translation quality. In spite of the fact that a lot of time is spent in explaining the logic behind the translation and how EC-SYSTRAN works, users often made remarks like “the style is very monotonous”, “what a stupid mistake” etc. However, there were others who really appreciated the facilities of modern technology and were truly thankful. This applied mostly to users that could now access the information written in a foreign language they did not speak. Furthermore, it should be noted that the users who rejected translations because of poor quality were very few.
The second year things ran more smoothly as a relatively steady number of users was reached, and this in spite of the frequent transfer of officials, which meant that the whole information procedure had to start all over again every time the head of a department changed. During the third year, most EC-SYSTRAN requests came from regular users, an additional person was hired to deal with the preparation of documents for translation (ie scanning, typing etc), and generally the activities of the office settled into a “routine” procedure. The development of the system is being continued according to the particular needs of the Greek public sector and users of the system are becoming more and more familiar with the concept of machine translation.
A happy ending? That will depend on whether or not the Office survives. Since the beginning of 1998, the whole project has been incorporated within a new EC framework (MLIS) and, at the time of writing, the proposal for an extension is about to be submitted to the EC. In any case, the Euromat office has definitely laid the foundation stone for the familiarization of Greek public servants with language engineering technology and its actual results. It encouraged people from public services to learn about how the recent advances in the field can help them in their everyday work, facilitating their activities. Opening people’s eyes to something they used to regard with prejudice and occasionally with fear, and transforming this prejudice into acceptance and enthusiasm has been a difficult but extremely rewarding task. This interaction between the developer and the end user has been the key aspect of this programme. Usually, developers and researchers work in isolation without ever seeing how users react to their products. Within the Euromat office, the opportunity has been presented for both developers and users to get in touch with each other. This way, developers got to know and “face” the users, while the latter had the chance to give their feedback and to enjoy the feeling of having contributed to a system they actually use. Each time a new, improved translation was given to a user, after his/her feedback and remarks had been incorporated, both groups —users and developers— enjoyed the results. Today, a wide range of enthusiastic users has been established, and their comments are most encouraging for the continuation of the work.
By Soula Fourla, and Olga Giannoutsou
Institute for Language and Speech Processing Athens