Maria Barros Eurescom barros@eurescom.eu

The European Union has 27 member states and 23 official languages, but multilingual websites are still rare in Europe. If the EU wants to live up to the motto “United in Diversity”, increasing the share of multilingual websites will be essential. This is particularly relevant for a “future Internet of contents, services and things, by and for people”, and when we are preparing the Web 3.0, in which computers are supposed to understand the users.

A Web in which computers understand the users’ requests and present concrete answers instead of a list of possible answers should be able to use multilingual content and retrieve multilingual answers. This would enable computers to search all available contents, independently of the language, and transmit the results to the user in the language that he/she understands. 

The challenge in Europe

The Web provides a unique possibility to build an information network for whole Europe, but the speed of the information society nowadays makes it difficult to share a European culture in which more than sixty languages are spoken and, among those, twenty three are official languages in at least one state. Even if Europe has a long history in translation and its economy is used to deal with a multilingual environment, making information accessible to everyone would require translation between five hundred six pairs of languages, once not all of the European citizens are multilingual or willing to learn other languages than their mother tongue.  

Global demand

Besides Europe, globalization creates an even bigger demand for a multilingual Web in the rest of the world, considering that out of an estimated 1.6 billion Internet users only about 464 million speak English, according to the latest Internet World Stats (31 March  2009). Access to online information is one of the big opportunities for society, and offering information for everyone is a major challenge that is far away from being solved. But, do we have the means to realise such a multilingual Web for a European Union united in diversity? 

Multilingual Web requirements

Web development and management for multilingual websites involve acquisition and annotation of language resources and a structural organization that allows automatic language translation. Even if fully translated documents are not required, contents should at least be prepared for machine reading and be available for multilingual search and retrieval tools. In order to address future barrier-free human-machine interfaces, websites must include multilingual recognition and synthesis, dialogue and translation. 

The situation today

Multilingual websites are already present in the Internet, but to what extent? It is not easy to manage a dynamic website in more than one language and it is certainly not easy to have new content available in other languages that are different from the original one. Multinational companies with a wide market view spend a lot of effort to keep their web-offer multilingual, but this involves a huge amount of resources and management. 

The project BabelWeb

Already in 2000, Eurescom project BabelWeb developed best practice guidelines for designing multilingual websites. The project developed a three-tier structure for the architecture of multilingual websites. Level one is a relational database for the contents; level two the overall structure, in which the contents are organised, and level three the presentation of the multilingual contents on the users screen. The translation of the contents was pointed as the crucial point for the planning of multilingual websites. We have today what is needed to implement the BabelWeb structure, with the advances of the language processing tools in the last decade. The general domain automatic translation came to a point where it is good enough to support human translations on static contents. We can already have a reasonable automatic translation for specific domains, e.g., e-health applications, specific commercial areas, political dialogs, law related services, etc…, and there are good results in multilingual search and retrieval systems to help finding contents in different languages. 

Conclusion

The multilingual Web is not only essential to the European Union but also for the worldwide information society, and the best tool for sharing knowledge and markets globalization. Its importance is confirmed by the European Commission’s efforts to promote multilingualism as part of the i2010 initiative to foster growth and jobs in the information society and media industries. A worldwide information space that offers content in every language and makes services available to everyone in their own language is a step towards a “Single European Information Space” in a multilingual Europe, where information access is no longer restricted by language barriers. 

Further information on Eurescom project BabelWeb is available at www.eurescom.eu/public/projects/P900-series/P923

Please send us your comments on this article.