Automatic language translation for user interfaces

Translations of user interfaces are an essential part of software development for the World Wide Web. Web pages are translated in three ways: by the developer as part of the normal life cycle, by the community that has a vested interest in the domain, and by machine translation. In this paper, we present an informal evaluation of the quality produced by these three approaches. We discuss the naive notion that language translations can be evaluated by doing to successive translations and evaluating the end result. The rubric used to evaluate the translations is presented.