The internet is no longer English only. The data is voluminous and the number of proficient linguists cannot match the day to day needs of several government agencies. Handling foreign languages is not limited to translating documents but goes beyond the journalistic written formats. Military, diplomatic and official interactions in the US and abroad require more than one or two foreign language skills. The CHALLENGE is both managing the user’s expectations and stimulating new areas for MT research and development.

The application of MT on the Internet has certainly attracted much attention in recent years, and many observers see its future mostly in this arena of real-time raw translation. However, the need for high-volume, fast turn-around translation of publication quality has not abated. This paper will take stock of that particular use of MT and venture predictions as to its future.

his paper is concerned with the technology of using the PARS English-Russian bi- directional machine translation systems in teaching English as a foreign language. This technology has no connection with the old form of computer-assisted language learning which uses «drill-and-practice» computer exercises and provides a sort of surrogate «electronic teacher». The main objective of the educational implication of PARS is to help the learner become familiar with the words in their normal contexts. The introduction of a machine translation system into teaching foreign languages is intended to get the most fruitful pedagogical results from the use of personal computers and expose the learners to the up-to-date information technologies.

ENGSPAN, a machine translation program (English-Spanish), has been used by the Translation Services unit of the Pan American Health Organization since 1985. In 1999, a total of 2,106,178 words were translated in that language combination, 86% of which were done with the help of ENGSPAN; the cost per word was 8.75 cents, that is, 31% below the normal rate. These positive results are explained by a combination of factors: the use of an MT program especially designed to meet the needs of the institution; the close collaboration of translators and computational linguists in the improvement of the program; the application of a pragmatic, flexible, and selective approach with regard to the quality of the end product; and in particular the support of competent translators who do the postediting work.

Our project Wired for Peace: Virtual Diplomacy in Northeast Asia (Http://www- neacd.ucsd.edu/) has as its main aim to provide policymakers and researchers of the U.S., China, Russia, Japan, and Korea with Internet based tools to allow for continuous communication on issues of the regional security and cooperation. Since the very beginning of the project, we have understood that Web-based translation between English and Asian languages would be one of the most necessary tools for successful development of the project. With this understanding, we have partnered with Systran (www.systransoft.com), one of the leaders in MT field, in order to develop Internet-based tools for both synchronous and asynchronous translation of texts and discussions. This submission is a report on a work in progress.

The Internet is a wonderful medium that frees its users from the confines of geographic boundaries. While the acceptance of the Internet is pervasive, the language barrier is somewhat tougher to overcome. Several options exists on the market to deliver multilingual content, few solutions can stand up to the dynamic demand of a modern website. Language context, translation turnaround times, and various business models are all barriers to creating a total solution for globalization and localization of websites. We will examine the difficulties in localizing a dynamic website and discuss the challenges we have overcome to create a dynamic translation platform.