Supporting Multilingual Collaboration for Wikipedia Translations

In Wikipedia, the largest encyclopedia on the Internet, a huge amount of knowledge is shared among users. However, differences in the number of articles among different language versions of Wikipedia represent an important issue. In order to solve the current imbalance of knowledge present in different languages , some users translate existing articles from one language to create new articles in another. These are known as " Wikipedia Translations. " Unfortunately, there are numerous instances of highly culture-specific articles in Wikipedia, which cannot be translated by individual translators. In such cases, collaborations among translators and experts are required. The author refers to this process as " Collaborative Translations. " For those who are not bilingual, it is difficult to explain culture-specific knowledge in languages other than their native tongue. The need for a multilingual environment arises which enables such users to communicate and collaborate with translators in their native language. In order to construct this kind of an environment, the author solves two problems: Designing a multilingual communication system for collaborative translations: The author will design a multilingual BBS system based on machine translations. Because of the features of Wikipedia Translations, the author has to consider two problems: First, the difficulty in using pre-defined languages for machine translation, since Wikipedia users with unspecified language abilities can collaborate on Wikipedia translations. Second, there can be many mistranslations of specific words in highly culture-specific articles that occur as a result of machine translators. Furthermore, the same specific words can be translated different in different articles. Controlling multilingual collaborations based on collaborative translation protocol: Wiki users are loosely coordinated volunteers, and they have various language skills. In the Wikipedia translation experiment, there were also several loosely coordinated participants. The author observed that users ii could not view progress of translation tasks and discussions in their collabora-tive translations because many discussions occurred in parallel and they could not follow all discussions. To resolve the first problem, the author designed a multilingual collaboration system on the Wikipedia system. In order to design this system, the author added multilingual abilities to LiquidThreads, a BBS running on the Wikipedia system. In order to solve the problem of difficulties arising from the use of pre-defined languages for machine translation, the author designed a system which translates messages in the viewing phase rather than in the posting phase. This system can answer viewing requests in any language. Moreover, the …