A Long Entry About The Future Of Wikipedia

Topic(s): me,computers
2007-08-10 02:06:48 PDT

The growth of English Wikipedia (in terms of number of articles) has grown at an exponential pace almost since its inception. It just a few years, English Wikipedia grew from literally nothing to one million articles during March of 2006. Eighteen months later English Wikipedia is on the cusp of two million articles!

One might use the exponential growth to predict that English Wikipedia will top three million articles mid-2008. But in just the last few months, an interesting trend has developed at the English Wikipedia: the article count dropped considerably below exponential growth.

There are any number of reasons why. The most obvious reason, at least for the English Wikipedia, is that there is a shortage of new subjects upon which to create new articles. Try to think of something very obscure. Search for it on Wikipedia. With near 100% confidence it'll be there. I'll call this "Wikipedia saturation."

Where that obvious reason is not true is on non-English Wikipedias. The German Wikipedia is quickly approaching one million articles, and several other Wikipedias are approaching one-half million articles. Their growth continues to keep exponential pace.

The obvious question to me is this: what happens when several Wikipedias reach their saturation point?

Let's say the English, German, French, Spanish, and Japanese Wikipedias all reach their saturation point.

I think at that point, Wikipedias volunteers ("Wikipedians") would attempt to create consistent translations of each article for each language. That is, no matter which language you're reading, the information and content would be equal for every article.

Now, that's a very hefty goal. Unattainable, really. Because each article on each Wikipedia is a living document. Changes to one document can't instantly be reflected on each language. Each language would need translators, and those translators would need to keep in sync with each other for every article.

But here's what's really interesting.

Suppose there is a subset of static (as in, non-changing) articles that have consistent translations across saturated Wikipedias. Let's say there are 1000 of them. Would it then be possible to create an intelligent automatic translator? A "Wikipedia translator bot"? I think it would be!

Such a bot would use the 1000 articles as reference data. The bot would automatically learn from those 1000 pages, and thus translate more pages. The bot's knowledge would grow upon itself to the point that all articles for saturated Wikipedias could be translated on the fly. That is, if some Wikipedian edits an English article, that edit would automatically and instantly be reflected in all saturated Wikipedias.

Let's go one step further. After the bot accumulates enough knowledge about the language structure of saturated Wikipedias, could it then translate non-saturated Wikipedias? Create an article where there previously wasn't an article? Maybe! It is possible.

At that point, every Wikipedia would have the benefit of every other Wikipedia. And then, at that point, Wikipedia would truly be the accumulation of every piece of knowledge on the planet!

name:
email:
url:
response: