PIVAX, an online contributive lexical database for heterogeneous MT systems using a lexical pivot

PIVAX is the first online contributive lexical database system allowing to create, maintain and manage the lexical resources of MT systems using a "lexical pivot", and heterogeneous because their language-specific components are developed at different places, with different linguistic approaches and computational tools. Only the most basic languagespecific lexical information must be stored in PIVAX, so that developers can use their own tools and protect proprietary information. An API is used for synchronizing PIVAX with other databases. PIVAX is written over the generic Jibiki platform. It is an evolution of the hypertextual PARAX multilingual database, and begins to be used by a U++C1 project using UNL UWs as lexical pivot.