Wikipedia Turns 10 and Sets Up in India to Grow Local Language Wikipedias
Wikipedia turns ten this January, and the encyclopedia that nobody expected to work is marking the anniversary by doing something it has never done before: opening an office outside the United States. The location is India, and the choice tells you a great deal about where the project believes its next decade of growth will come from.
The groundwork was laid through 2010. Senior figures from the Wikimedia Foundation made repeated visits to India, held public meet-ups with the local community of editors, and appointed the first Indian to sit on the Foundation board, Bishakha Datta. A formal Indian chapter has been in preparation for some time. The offshore office is the concrete result of all that travelling.
Why India
Two reasons, and only one of them is about numbers. The first is the vast pool of potential internet users, a population coming online at a pace few countries can match. The second, less obvious but arguably more important, is ground support: India already has a passionate community of Wikipedians who edit, argue, organise meet-ups and recruit newcomers without being asked.
Wikimedia has learned that money and servers do not create a functioning encyclopedia. People do. Trying to seed a community from scratch in a country with no existing volunteer base is slow and often fails. India offered the rare combination of enormous headroom for growth and an existing core of editors to build around.
The Indian internet is not one thing
The Foundation has been careful to say what it learned quickly on the ground: the Indian internet is not a homogeneous entity. It is a patchwork of languages, scripts, connection speeds and devices. English is widely used in Indian business and higher education, but it is a first language for a small minority. The country recognises more than twenty official languages, and hundreds more are spoken daily.
During his visits, Wikipedia co-founder Jimmy Wales returned again and again to the same point: growth in India has to be approached from a strictly localised perspective, by expanding the user base for the local language Wikipedias rather than pushing more people towards the English one.
That distinction matters more than it sounds. English Wikipedia has millions of articles. Hindi, Tamil, Bengali and Marathi editions have a fraction of that, written and maintained by far smaller groups of volunteers. A reader who cannot comfortably read English encounters an encyclopedia that is technically available and practically empty.
Local language editions are not translations
There is a common misconception that a language edition of Wikipedia is a translated copy of the English one. It is not. Each edition is an independent project with its own editors, its own editorial policies and its own idea of what deserves an article. Some of the content is translated; much of it is written from scratch about subjects the English edition barely covers.
This is the same lesson companies learn when they take a website into a new market. Running text through a translation engine produces something readable and useless. What works is adaptation, and the businesses that do it well treat localization services as a rewrite for a different audience, not a word swap. Wikipedia stumbled into that conclusion years before most corporations did.
The obstacles are practical
Anyone editing in an Indian language faces problems an English-speaking editor never thinks about.
- Input. Typing in Devanagari, Tamil or Bengali script on hardware designed for the Latin alphabet is awkward, and keyboard support has historically been poor.
- Sources. Wikipedia demands citations to published sources. Material in regional languages is often not digitised, and offline sources are hard for other editors to verify.
- Fonts and rendering. Indic scripts have rendered badly on cheap devices for years, and text that displays as boxes does not attract readers.
- Volunteer numbers. Small editions are fragile. A handful of active editors can carry a language edition, and a handful of departures can stall it.
Ten years of an unlikely idea
It is easy to forget how implausible Wikipedia sounded in 2001. An encyclopedia anyone could edit, with no paid writers and no advertising, was widely predicted to collapse into vandalism within months. Instead it became one of the most visited sites on the internet and the default first stop for hundreds of millions of people.
The tenth anniversary finds the project in a curious position. The English edition is mature, heavily policed and growing slowly. Growth now comes from everywhere else, which is why the Wikimedia Foundation is spending money on offices and outreach rather than servers. Editions such as Hindi Wikipedia are where the next hundred million readers will actually arrive.
What success would look like
Not article counts. Anyone can inflate those with bot-generated stubs, and several language editions have embarrassed themselves doing exactly that. The metric that matters is active editors: people who show up next month, and the month after, to fix a fact or add a paragraph.
If the India office works, it will be because it created editors, not articles. That is a slow, unglamorous business of workshops, university partnerships and patient explanation of citation rules. Ten years in, Wikipedia has learned that this is the only thing that has ever worked.
A template for the rest of the world
India is a test case, and the Foundation knows it. Brazil, Indonesia, Nigeria and the Arabic-speaking world present the same shape of problem: large populations coming online fast, thin encyclopedic coverage in the languages people actually read, and small volunteer communities carrying an unreasonable load. Whatever works in Pune or Bengaluru will be copied, and whatever fails will be quietly abandoned. For a project that spent its first decade being told it could not possibly work, that is a familiar way to proceed.