Brownstone Design in Cambridge

Log in
updated 12:28 AM BST, Apr 23, 2014
Technology and life sciences news from the Cambridge cluster

Running Wikipedia, possibly not as easy as Jimmy Wales makes it look

  • Written by Lautaro Vargas
  • Published in Blog

Jimmy WalesWhen Wikipedia emerged from the rubble of the dot.com crash at the beginning of the millennium, it was a refreshing reminder of the aspirations many first held for the internet, a tool for universal good, connecting the world and offering free information that was free of pop-up ads.

The online encyclopedia grew to become one of the 10 most popular web sites in the world (pdf) with 408m visitors to the site according to ComScore (founder, Jimmy Wales says its own servers suggest it's probably closer to 600m).

It is closing in on 20 million articles in 270 languages, 200 of which have at least 1,000 articles. All of this driven so far by a staff of just 70, though backed by around 100,000 volunteer contributors.

Great job Jimmy, well done, easy. Except it wasn't of course. What struck after Wales finished speaking to a packed hall last Thursday in the first of Cambridge Network's new lecture series, was just how simple he had just made it all seem, you almost felt tricked as you realised 'only 70 staff, 20 million articles, 270 languages... hang on.'

Part of the deception lay in the straightforward mantra that underpins the work of Wikipedia and the Wikimedia Foundation behind it:

'Imagine a world in which every single person on the planet is given free access to the sum of all human knowledge.'

But it's in striving to achieve this where Wales says Wikpedia's biggest challenges lie.

Broken down, it means getting internet access to the most remote places in the world, it means getting people to post articles in remote languages, it means convincing governments not to restrict access to articles and it means to wrestle editorship away from the almost exclusive grip of the twenty-something male geek.

Wales wants at least 250,000 articles in every language of the world that has at least one million native speakers, that's 370 languages, a tough ask for any organisation, let alone one with just 70 staff.

For instance, in Liberia one university has a 256kb internet connection for 1,000 students for which it pays $3,500 a month. It gets around the need for a decent web connection for Wikipedia by using a 2 terabyte box on a server which carries all the Wikipedia information as well as relevant journals. It is a very useful model that can be replicated elsewhere, but it doesn't provide an answer for aspiring contributors.

Then there's how to cater for all the languages with existing technologies. Some of the many Indian languages struggle with the English language keyboards so often used. One answer would be automatic translation programmes, but they are barely up to the task when it comes to the world's most common languages, let alone the various Indian languages.

In Africa it's even harder, there's only one indigenous language with more than 1,000 entries, Swahili.

And it isn't just about the developing world, here, in the Western hemisphere Wikipedia suffers from a lack of diversity. About 87 per cent of contributors are male, 26 years old and have double the percentage of PhDs compared to the general public (they're very very geeky says Wales).

This has an impact on entries, for instance people repeatedly tried to have an article on Kate Middleton's wedding dress which could be a significant article in terms of the fashion world deleted said Wales. Meanwhile there's over 100 Linux distribution articles.

Wales wants at least 25 per cent of contributors to be women and he wants the average age to rise. Part of the challenge is to make the editing interface more accessible and faster, at the moment a lot of the editors have a lot of spare time.

To do this, all of this, takes major resources, from the basic tech support to the outreach to the interface overhaul, hardware costs, domain names, site hostings and as we know Wikipedia and all its sister projects are entirely reliant on donations, there's not even any advertising (except for donations).

Last year's budget was just over $20.4m, the most it has ever raised. Next year it should be around $28m, but still, that's not a lot to be working with, less than £70,000 for the 370 language target, but it's all they have for now.

Wales used to work in one of the world's most notorious futures and options trading floors, the Pit in Chicago, another tough environment. I bet he made that look easy too.