The IAMT Certification initiative and defining translation system categories

1. Background Since the middle of the 1990s there has been a rapid increase in the number and variety of translation systems available, in the form of stand-alone software for 'automatic' translation, computer-aided translation systems for large corporations, translator workbenches, translation memory systems, on-line systems provided on the Internet (some of them free), and there will no doubt be more in the future. For the general public, computer software for translation is a quite new product; they are unaware of the advantages, limitations and methods of using such systems. They are furthermore familiar with rapid improvements of computer technology and software, and will therefore be expecting similar rapid improvements in the quality of translation software. One of the tasks of the MT community must therefore be to convey some idea of the immense complexities involved in dealing with natural language (and in particular the kind of colloquial language used over the Internet), and to explain why automatic translation has progressed so slowly despite over forty years of intensive research. For some time the IAMT has been concerned that some guidance should be provided for potential purchasers of systems, to explain, for example, the differences between MT and translation support systems, the differences between systems for 'home use' and systems for large companies, etc. The guidance should also cover the suitability of different types of systems for particular uses and tasks, information about components of systems and how they may be used (including their benefits and limitations), and information for the cost-effective and/or time-saving use of different types of translation support and MT systems. In addition, it is recognized that ideally there should be some authoritative benchmarking or some reliable consumer evaluations, but these activities are currently beyond the resources and competencies of IAMT and its regional associations. It has therefore been decided that as a first step IAMT should attempt to provide a categorization of translation systems together with explanations of what different types of systems can and cannot do. The IAMT initiative began in 1997 with the Council setting up a small subcommittee under Eduard Hovy (president of AMTA). Initially the idea was that IAMT might establish a »seal of approval« which would be awarded to companies conforming to agreed standards. However, it was quickly realised that this term would imply some evaluation of systems, and this was not intended. Consequently, it was decided that given the limited resources …