article taken from http://www.guardian.co.uk/technology/2009/mar/26/internet-translation-project
Jim Giles guardian.co.uk , Thursday 26 March 2009 16.37 GMT "A new Wikipedea-style project will rely on users to upload translations of blogs and newspaper articles English translations of Arabic newspapers could soon be online. Could online language barriers be overcome using the same crowd-sourced approach that has made Wikipedia a success? That's the aim of the Worldwide Lexicon (WWL) project , which released a suite of new translation tools last month. Visitors to any blog or newspaper that uses the software will be able to click on to the site's text and enter a translation in the language of their choice, just as Wikipedia users can edit any entry in the encyclopaedia. The translated version will then be accessible to other visitors. "Language is one of the few remaining barriers on the internet ," wrote Brian McConnell, the San Francisco-based programmer behind the project, at the time. "The web has rendered time and distance largely irrelevant, but much of it remains fragmented by language." English is the online lingua franca, yet it is the first language for only 6% of the world's population. And even though English dominates the internet, Chinese and Japanese content accounts for around a fifth. The Arab-speaking world, which arrived online a little later than Europe and the US, has also been one of the fasting-growing portions of the web in recent years. McConnell says that bilingual readers would be prepared to provide free translation services between these and other languages if they were passionate about the content involved. Speakers of around 15 different languages, including Spanish and Japanese, have already used an earlier version of McConnell's tools to translate posts from Boing Boing and Make blogs. Fan sites might also find willing translators – some English football teams have large followings in Asia, for example. The success of Wikipedia shows that high-quality content can be created using volunteers, provided the subject matter attracts enough interest. But Mary Margaret O'Hara of ProZ , a website that freelance translators use to find work and collaborate on projects, says it would be difficult to control quality – particularly if a mistake were to filter through several different translators. "This could eventually garble the message," says O'Hara. "There is no way a website owner could check if they don't speak the language. It's a great deal of trust to place in another person." McConnell says that website owners could configure the WWL system to guard against problems like that. Unofficial fan sites, such as those that track sports teams or videogames, might be happy to let anyone translate, but if quality control was a priority, content providers could restrict access to a few trusted, or paid, translators. Translations can also be rated by other users. Some sites might be happy to let anyone work on their content, but reserve publication of the translations until a high-rated user has checked it. Even with some degree of checking in place, the service could cause legal problems if a mistranslation were to end up producing a libelous statement. But the transparent nature of the WWL service should deal with that, says McConnell. Users can be banned and their IP addresses traced, just as sometimes occurs on Wikipedia. Site owners can also use disclaimers to make the legal status of the content clear. One site that is planning on using the tools in Meedan, a San Francisco-based social networking service that attempts to bring together Arabic and English speakers. The site, which is currently in the final stages of testing, is built around discussions of Arabic and English blog posts and news articles about the Middle East. Translations are done by machine and then edited by interested users. Ed Bice, the site's founder, says that Meedan is building a Firefox plugin based on the WWL system. The plugin will allow users to translate an article and simultaneously upload it to Meedan. This should lead to more content on the site and higher quality translations, says Bice. McConnell hopes to make money by tailoring the system to the needs of publishers and says he is in discussion with a "major" business publication. He also wants other developers to build on the system. If a website owner wants all their content to be translated immediately, for example, they could use the WWL to link to machine translation services such as Google Translate. The machine-translated version could then be later improved by users. McConnell has already bolted automated systems for translating Twitter and RSS feeds on to the WWL framework. But many language pairs, such as Hindi-Urdu, are missing from the major automated services. If translation engines for those pairs became available they could easily be added to the WWL, says McConnell. " man, with so much of posting updates and news, i'm starting to feel like an rss to email feed translator, or a posting robot. * grins * with regards, ruchir. -- As long as forever, I will stay by your side, I'll be your companion, Your friend and your guide!!! www.ruchir89.wordpress.com To unsubscribe send a message to [email protected] with the subject unsubscribe. To change your subscription to digest mode or make any other changes, please visit the list home page at http://accessindia.org.in/mailman/listinfo/accessindia_accessindia.org.in
