Research intelligence - What have you been missing?

Another blow has been dealt to the foundations of the Tower of Babel with the addition of an automatic translation service to the World Wide Science Alliance's online search engine. The facility will allow scientists to translate open-access research - mostly carried out by government bodies on topics such as energy, medicine, agriculture and the environment - from, or into, nine major world languages.

WorldWideScience.org was launched in 2007 by the British Library and the US Department of Energy's Office of Scientific and Technical Information to provide a single portal - www.worldwidescience.org - to carry out federated searches of multiple open-access scientific databases around the world.

The alliance now encompasses about a dozen national libraries and research organisations, and the database links to 400 million pages of research in 65 countries. The search engine has expanded its users from fewer than 8,000 in January 2009 to nearly 96,000 in April 2010, while the languages employed by users have risen in number from 71 to 114.

The translation tool was rolled out at the International Council for Scientific and Technical Information's annual conference in Helsinki earlier this month.

Richard Boulderstone, director of electronic and information systems at the British Library and chairman of World Wide Science Alliance, said that although most scientific papers are published in English, governments around the world are increasingly carrying out research of their own that is not translated and would not otherwise be accessible to researchers who do not speak the country's language.

He said it was difficult to say how valuable such information might be to researchers in other countries.

"Without a translation tool, we simply do not know what we are missing out on," Mr Boulderstone said.

"Translation will break down the language barrier and enable non-English-speaking researchers to access our research and vice versa, to enable greater levels of knowledge sharing and global cooperation on future research."

He said it would also reduce the risk of research being unnecessarily duplicated, especially on important global issues such as climate change and food stocks.

"Government reports are pretty hard to find because they go up on government departments' websites, and people don't know they are there," Mr Boulderstone said. "Our site is a way of creating a search of those sites simultaneously."

The new multilingual beta site, which uses Microsoft Research's translation technology, will allow real-time searching and translation into English, Chinese, Japanese, Korean, French, German, Spanish, Portuguese and Russian.

Mr Boulderstone is hopeful that word of mouth and presentations at conferences will attract users from even more countries. He would like to add more languages to the translation tool as part of its ongoing development.

Because of its small budget, the translation project was possible only with the help of Microsoft, which donated the technology and hopes to showcase it via the site.

Language is no barrier

Mr Boulderstone admitted that automatic translation could be clumsy, but he insisted that it has improved a lot in recent years. "It is not spoken English, but you can make sense of it. We will listen to the feedback we get from users. Imperfect translation is better than having documents only in Chinese. If a document looks interesting, researchers can get it professionally translated."

Keith Jeffery, director of IT and international strategy at the Science and Technology Facilities Council, a user of WorldWideScience.org, agreed that automatic translations serve a purpose. "They are a bit word-for-word rather than being grammatically perfect, but you get the sense of it. They would not be adequate for people studying literature, but as a scientist you are grateful for any translation."

He said repository search engines such as WorldWideScience.org and OAIster, which also searches open-access databases, are useful complements to Google because they could find resources that are not on the World Wide Web.

"Google only scuttles over the surface of the web, but WorldWideScience.org and OAIster allow you to find deposits of interesting minerals below the surface," he said.

Professor Jeffery said the translation tool is an invaluable addition to WorldWideScience because it opens up a "huge amount" of non-English material to English-speaking researchers, especially from China, South Korea, Japan and Russia.

"The rise of China makes this development incredibly important for the future," he said. "There are already more web pages in Chinese than English: it is where things are going."

paul.jump@tsleducation.com