At the Internet Governance Forum being held in Hyderabad, Ajit Balakrishnan, CEO of Rediff.com said that there is no evidence from the last ten years of the Internet business that users want Indian languages. Rediff has email in 11 languages, and 99% of the users prefer to use email in English. One of the issues is that “practically all of the 300 million young people who aspire to something in this country aspire to learn English.” Therefore “Let us not assume that users want Indian languages.” He mentioned that Nokia has experimented with Indic language keyboards, and pointed out Eterno’s transliteration app which allows the usage of latin characters for messaging in Indic languages.
During the Q&A, Ram Mohan from the audience put forth a significant point on the requirement of multilingual standards – for the creation of a common set of semantics and terminologies, and the need for a framework and a common structure for script and language-based solutions. “We’re talking about a problem that begins at the core of the Internet, at the domain name system, and goes all the way to Internet navigation. Multilingualism is often confused with Internationalized Domain Names. One is not the same as the other.”
He suggested that a standard or a shared model for the adoption of scripts and languages online be created. “…people think languages, but computers work with scripts. And we need to find some common way to bridge that gap. Otherwise, I think we run a real risk of having simply scripts depicted online and not having languages. Some languages may completely miss this transition from an oral world to a digital computer-based world.”
Responding to him Balakrishnan said that one may sit in a committee and madate people to do what you want, but they will do things which are more convenient to them. Which is why TCP/IP won over the X400. “You find all this energy is being spent on local scripts as URLs for domain names, I encourage you to think about it at all.”
Balakrishnan doesn’t believe that the Internet is about content – most young people are using the Internet to send messages, download music, view pictures or videos. None of this is particularly language related, and virtually 90% of the content is not text based. He believes that the PC era ended last year, and the future of access is mobile, but it wont be text based mobile.
“People are frantically working to master the voice-to-text conversion piece, not because we want to convert it for any other reason but for the fact that most of the text, most of the processing software has historically been built around text processing, indexing methods, TDIDF calculations and so on. So the big thing, the really big thing over the next five years – whoever gets a breakthrough on a voice-based Internet where you can speak into it and hear things back, that is a big price. If that happens, you will find all your conferences have been wasted.”
“Voice-based Internet is where the future lies. And if national entities have to be pushed to do anything, it’s to make sure you make the voice-to-text recognition system accurate. At the moment in India we are not getting results more than 70% accuracy. If you can use the brains and get it to 95%, I think that is fantastic. That will solve all our problems.”
Update: BG Mahesh, CEO of OneIndia, has commented on his blog, differentiating between reading and writing content, saying that the time has not arrived for the usage of Indian language in email. Some data:
During the initial years about 80-90% of our language traffic was from outside India (NRIs). Post 2006, 60% of our language traffic is from India (I am averaging across all languages). The readership in India has grown because of the growth in Internet user base (40m or 60m – whatever).
Note: Do check our Indic Language section.