Date:24/01/2010 URL: http://www.thehindu.com/2010/01/24/stories/2010012460261200.htm _________________________________________________________________________________
Link: National Pushing the frontiers of transliteration T. Ramachandran It gains ground because "the Indic web is now finally coming of age" _________________________________________________________________________________ Indian languages form a small percentage of total traffic of Google Creation of content in local languages rising significantly _________________________________________________________________________________ KOCHI: "Typing in Indian languages has never been easier," says the website of the popular blog publishing tool Blogger. For, English can now be used to easily generate text in a number of Indian languages, without the user having to grapple with the intricacies of an Indian language keyboard. The software does the job: it transforms the English 'input' of the particular language (based on pronunciation) into the corresponding script. For example, a user types in hamara (in English), and the software comes up with the Hindi equivalent. Called transliteration, it is being increasingly used online these days. Transliteration has been gaining ground because "the Indic web is now finally coming of age," says Rahul Roy-Chowdhury, product manager, Google India. In Google's case, "across Gmail, search and all of the properties, Indian languages are still a small percentage of the total traffic. But the growth rates have been much, much higher than English," he said in a telephone interview to The Hindu . He sees the creation of content in local languages rising significantly, "across a much wider spectrum of topics." "The Indic web will become a lot more mature." Smart tools It is also becoming easier to do transliteration; tools have become smarter and easier to use and almost universal. Google, for instance, launched its input method editor (IME) in December ( Link: http://www.google.com/ime/transliteration ). Once downloaded and installed, it adds transliteration capability to the existing word processing programmes on the user's computer, like Notepad or Microsoft Word, or other programmes that support such input. The interface options it provides can be used to generate text in Bengali, Gujarati, Hindi, Kannada, Malayalam, Marathi, Punjabi, Tamil, Telugu and Urdu. But transliteration tools do mostly come with an independent Web interface (example: Google Link: http://www.google.com/transliterate). Over the years, Google has also been incorporating transliteration into its web-based services such as Gmail and Orkut, apart from Blogger. It also offers a few other transliteration tools and options. Microsoft recently launched the web and desktop (beta) versions of its Indic Language Input Tool (http://specials.msn.co.in/ilit). It offers transliteration support for Bengali, Hindi, Kannada, Malayalam, Tamil and Telugu. The development of transliteration tools has been on for a long time, though. Asked what prompted him to come up with the transliteration tool, ITRANS, in the 1990's, its developer Avinash Chopde explained: "As a graduate student in the U.S., I felt it would be very easy to use the current technology to create tools that would allow easy use of the English keyboard to produce Hindi output." He no longer works on it; the new generation tools are "very good." The earlier systems were somewhat cumbersome; users had to be fairly technically savvy to make these work, explains Mr. Roy-Chowdhury. So what Google did differently was to develop "extensive machine learning algorithms and data-mining tools at the back end" to make intuitive transliteration possible. Quillpad, another prominent player in the transliteration arena, states on its website that its tools are 'intelligent' and can largely guess the user's intent when an input is made ( Link: http://www.quillpad.in). Sowmya V. B., an IT professional and blogger, says such tools make fewer mistakes and offer the user a list of word options with the best choices when they are unable to properly decipher input. Zero learning And there is also the question: should not more attention be paid to develop better Indian language keyboards? The problem with such keyboards is the learning curve involved. "With transliteration, it's more or less zero learning. However, for people who are not used to typing in English, I think keyboards are the best way... I don't use an Indian language keyboard, but people who use it say you can type faster once you get used to it," says Sowmya. 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
