Otelo and the Domino translation object

MT Summit VII Sept. 1999 Otelo and the Domino Translation Object Alan Barrett Lotus Development Ireland Abstract Lotus is working with SAP and a number of MT vendors to make usage of MT easier. Most of this work has been done in the frame- work of the Otelo project. It is also part of Lotus' efforts to make development of multi- lingual web applications much simpler. Terminology interchange and text interchange formats as well as a Linguistic Services API are discussed. Also covered is the Domino Translation Object which enables use of these technologies on the Domino infrastructure. 1 Introduction The lack of access to knowledge because it is only available in languages not understood by people who need that knowledge is a huge economic issue. This applies to knowledge on the web, knowledge needed to effectively transact e-business and to knowledge of wider cultural value. This is particularly an issue for people who only understand "minority" languages. While many senior people in organizations may under- stand a range of "international" languages, people at a lower level in organizations may be excluded from full participation. Existing and potential customers or business partners may also be excluded. Lotus is working with SAP and a number of MT vendors to make usage of MT easier. Most of this work has been done in the framework of the Otelo project which has been part funded by the European Commission. However, this paper is not a full discus- sion of the Otelo project. Four contributions to MT quality and access to MT are discussed in this paper:- • OLIF (Open Lexicon Interchange Format) • OText (Open Text interchange format) • Linguistic Services API • Domino Translation Object 2 OLIF In order to make MT more useful to organizations there needs to be a simple way of importing terminol- ogy into the proprietary lexicons of MT engines. A corporation may have a large store of terminology available (e.g. from human translation efforts) which, if easily usable, will make MT much more productive. Additionally, if a corporation has control of it's termi- nology in it's own repository and it can easily be imported by MT engines then the corporation will probably be much more comfortable with MT (i.e. the corporation's terminology will not be locked into a proprietary format). It will also allow the corporation to use the best MT engine (along with the MT engines own supplied terminology) for any particular job. Open access to terminology preserves an organiza- tion's expensive investments in terminology development. OLIF is a standard XML interchange format open to all. An entry in OLIF consists of: • Central entry defined on a concept basis (single & multiword units): defining features are canonical form, syntactic category, domain classification/reading & language • Linguistic description (for each stem variant and also for the entry as a whole) based on what MT systems really use • Links to other entries such as synonyms and transfers to other languages (including equivalence type and additional tests …

Content truncated to 500 words