Home » , , , , , , , ,

#NAMAindic: Why and how Indic languages can thrive on the Internet

Share on Facebook0Tweet about this on TwitterShare on LinkedIn8Email this to someone

This January, MediaNama held an open house discussion on the future of Indic languages online. This is Part 3 of our coverage of the discussion. Read Part 1 here, and Part 2 here.

E-commerce and advertising players have been slow to take up Indic languages, even with the huge boost of Indic users that they received post-demonetization. What can be done, therefore, to improve the adoption of Indic languages online, and to bring the quality of Indic user experiences at par with English?

No half measures: Dr Govind, who ran NIXI for many years, said: “I feel that half-baked solutions won’t work in these things. Having some content in one language and some content in another language won’t work. We need to work out complete language solutions end-to-end in languages like Hindi, Tamil and Malayalam.

“There are many piecemeal solutions. There are no complete solutions for OCR, machine translations. It needs to be perfect, because unless you throw it to the public and work out the various words, the things which are required in the contextual way, the English to Hindi or Malayalam translations — unless we work out in a total solution way, things won’t work out easily for Indic languages. The other thing is that we have to use the languages whatever way it is there, like .bharat domain names are there in Hindi and various language names, so why not use these domain names? Like [Ajay] Data was saying, there’s already an email solution. If we start making the SME sector aware of this technology, they’ll still benefit even if larger companies and payment solutions can’t.


“I think that apps and emails are different. In emails you can get way more information and there are connecting IDs. In apps you can’t have WhatsApp kind of things. I think when Rediff was trying out local language emails, the ecosystem wasn’t there. But now we have a local language ecosystem that would allow these things to come up.”

Coercive action: Rakesh Kapoor from Process 9 put things plainly: “Are we in 1995 or 2017? All smartphones in this country should be 100% localized. Everything that comes bundled in a phone should, therefore, be available in all Indian languages. Telecom content must be available in local languages. Even in e-governance, there has to be a mandate that there can’t be a government site or app in this country which is not localized in all languages. Payments: there has to be a mandate that you can’t have a payment app in this country if it is not localized in Indian languages. If we want 85% of this country which does not speak English, to be empowered, all content can’t be in English. Do we have a will, whether it’s Mr Narendra Modi himself, or whosoever, do we have a will to say ‘this is how India is going to go forward, and you either fall in line or get out’?”

Pay for quality, get quality: Tanmoy Goswami, a journalist, highlighted that Indic content is often sub-par because news organizations don’t invest in it adequately: “I think it’s all very good to talk about becoming savvy and investing and doing better content so that you can get a better audience onto your platforms. But if you pay peanuts to your Indic language journalists, you’re going to get monkey shit. In 2014, when I was writing a story for Fortune on the state of native language media in India, I spoke to 25-odd journalists from various languages, Marathi included. Marathi, Bengali, Hindi, Gujarati — and all of them said that even the editors in their newsrooms get paid maybe one-third of what their English peer would fetch. Again, this is a chicken and egg situation.

“The management says, these are not moneymaking businesses, so why should we pay more? […]And when I asked them why that was the case, at least one executive from a company told me very coolly that they’d rather hire an English writer, get it translated into Marathi, and pay the translator, and it would end up being cheaper for them than to hire a Marathi tech journalist. So I think wish number one is: start paying, start respecting your staff a little more. Second thing, I think Vinay Parab who used to be the editor of Loksatta, in the 1980s when tech literacy in India was low, Loksatta was one of the first papers to start demystifying technical jargon for its audience. And Vinayak was the guy who used the word “Pranali” to explain operating systems. He did a 90-part series on demystifying jargon — things like Cathode Ray Tube etc.

“Initially there was resistance from the readership, but because at that time the newsroom culture gave respect to writers, they persisted with that kind of content, and after that it became a blockbuster, that column. You start paying your people more, boost their morale, they will feel more invested in producing better content, you will be more in-tune with what’s happening in the larger journalistic fraternity, and that’s how you will probably salvage your brand reputation.”

Getting more Indic news content: Divyashree Bhat, Program Manager at Google’s Indic initiatives, said: “My part of the wishlist would be to have a breadth of news content. If I’m someone who’s just looking for lifestyle, health, education content, that’s largely a part of the long-tail content that I consume in English. Do I find that in Indian languages? I’m a Kannadiga. I know certain newspapers that I would go to and consume content, then I have certain local e-papers I go to. I consume content about my home town. But beyond that, I’m very limited in that perspective, right? Where I know that this is what my newspaper puts out, and this is what I consume. But when I look at English, if I want to just take a look at probably just gadgets, or for that matter travel or policy, I have so much more to choose from.

“I think that is something that’s very much on my wishlist: can I have a larger breadth of content? Indic news publishers have done a phenomenal job of getting a lot of content online. But activating that longtail and getting that breadth of content would definitely be on my wishlist. Secondly, coming to search, I think that’s the wishlist for search as well. Your search results always get better with a lot more breadth in content. That is also something that I think will help more. What we see is definitely the role of input mechanisms in search. One is a wishlist of the kind of content we want to consume, but the second is: how do we enable Indic-first users? We are all probably prefer Indian languages, and we’ll largely be English-first. We’ll probably test out a few apps in Indian languages, consume content occasionally, but we’ll be English-first. But what about the Indic-first users? Are we empowering them with enough input methods? We have enough keyboards, we have handwriting input, voice… how do we improve that? I think this is a personal challenge — how do we improve that input mechanism? Can we make it easy and simple?”

Getting Indian scripts to the DNS root zone: Samiran Gupta, head of India at ICANN, said: “We need to remember how the Internet works. The Internet is a system that actually helps you to retrieve data or information that is sitting somewhere. And the way it works is through the IP and DNS systems. Almost 95% of the Domain Name System (DNS) is in the Latin script. A lot of references have come, for example we have talked about Indic-first users. So do they have to learn English to then get content in Kannada, or can they actually get there [without doing that]? Until we have Indic language DNS, you will not have competition or choice. I sat through a couple panels earlier today, and a gentleman mentioned that the way it came out was that they were forced to do Indic languages from English because of demonetization.

“The point simply being, and Tanmay talked about Indic journalists not being at par — the point really is, if we keep the content aside for a moment, let’s just talk about accessibility to the content. There is a process going on right now — most Indic content, or scripts rather, are housed under something called the Neo-Brahmi scripts, there are 9 broad scripts in that palette. Essentially, if we can get the Unicode versions of those scripts created, and then put into the root zone, the searchability will improve dramatically. A community-led effort is on and like anything community-led, it starts off, it flats off — we kickstarted it again, and folks from Reverie, I’m short of putting a gun to their head — Mr Ajay Data was here earlier.

“I want them to come on board as well, because the quicker we can capture the Indic languages into the DNS, the faster there will be a massive change. A lot of the questions you’re asking, for example, should we use an email or pushing text is good enough? Or should we use voice? Until you have a DNS which runs on Indic script, which now runs on Latin and other scripts including Chinese, Cyrillic, Japanese and Korean, you won’t get the depth or coverage. I just wanted to make this statement, invite folks in the room if they’re interested to join the panel, I’d be happy to facilitate that, because that panel is working towards taking eventually 22 scripts to the root zone.”

Share on Facebook0Tweet about this on TwitterShare on LinkedIn8Email this to someone