Monday, January 12, 2009

Online crowdsourced translation

I was jumping between videos on Youtube (Jatropha yields 6 times as much usable biofuel precursor oil as soy or corn, AirNZ did a testflight :) when I found a video from a Korean Mushroom grower. It had a voiceover competing with machine noise, but this was spoken by a young woman with.. pretty fluent English, not perfect, but quite good. I found myself thinking 'What if these people and others like them had a good, free(?) translation service online?' Would they be more likely to do business with English speaking customers or business partners? Could the reverse be true?

So here's the idea: Online crowdsourced translation.

Like Wikipedia, it utilises the 'wisdom of the crowds' - not to record all human knowledge, but to enhance language skills and improve communication between different cultures and countries.

It could be free and wiki-based to keep the individual effort needed minimal.
It could have a reputation system, so that prolific, competent translators could be lauded.
It could charge a fee for fast translations, or for privacy (keeping the translation effort within a group of high reputation contributors). If it was to be as charitable as Wikipedia, this money should be channeling towards upkeep and enhancement of the service.

The value proposition for people needing translations is clear - like Wikipedia, they go there for better information. The value proposition for contributors is less clear. Maybe they could go there and perform some minimal effort for charity. I edit Wikipedia for charitable purposes, but I go there initially because I want some information - I want to profit from some of that charity myself.
Perhaps they could do it for monetary reward. Though I think the service would work best if it were at least primarily free. Perhaps if paid, fast, private translations were offered first to those with a high reputation on the site (a reputation gained by first performing free translations)?
They could do it in the expectation that one day they themselves would need a free translation.
They could do it to learn another language better.
They could do it to make friends with people from other cultures (remember, weak social links are powerful).
They could do it to learn what other cultures are interesting in right now.

Throughout this thought experiment I've been thinking not of translating straight from one language to another, but just of improving the spelling and grammar of work already translated by people with less than fluent language skills. The former is more useful (and is already done pretty well by software) - the latter is more amenable to the 'wisdom of the crowds' approach. After all, more people speak a language fluently and can understand less than perfect use of that language than can speak more than one language fluently.

One last thought: This will be a useless service eventually, as software translation gets better. But it could make a nice improvement to international relations and trade in the short term, and could perhaps be adapted in the future to more specific knowledge, such as locally correct spelling and dialectic words (what is the most accurate transation of the New Zealand 'Sweet as, bro' to an audience in California? in Ontario? in St. Petersburg?).

There's certainly a large audience of people who would like a better shop sign for tourists or an international press release, or just to learn another language that little bit better. Markets are global and China and India are growing fast.

No comments: