Google Translation Center

Google took the lid off a new service today, an extension of its Translate service called Google Translation Center that connects translators with people who need content translated into other languages. All compensation arrangements are left to the individuals involved, but Google will store results on its own servers.

Already the new service has been compared to Amazon's Mechanical Turk service, which matches up people/companies needing services with those who can provide them. Google doesn't take a cut of whatever arrangements are made, leaving lots of speculation as to how the new service fits in with the search company's overall strategy. More on that later.

A little detective work at Google Blogoscoped blew the whistle on the service perhaps before Google intended: Philipp Lenssen's thorough analysis of what the Translation Center suggests the service was briefly online before disappearing. Subsequent posting around the Net may have prompted it into a permanently live status.

The idea is pretty simple: Those in need of translated documents can browse translators and work out the details with them, and can submit the material via the Translation Center and post a request. Translators can post their services and make use of Google's new "easy-to-use translation tools."

Google cites not only professionals, but also volunteers as potential sources, and the service will match current translations with previous ones to prevent duplication. The value is fairly obvious to those in the translation business and to those, like webmasters, academics, etc., who need content translated into as many as 40 languages (Google Translate's current capability).

But what about the value to Google? Investors have criticized Google and other companies for spending too much time on peripheral services that seem to do nothing to enhance their core business, which is search advertising. Google Translation Center, then, seems to fit into that peripheral services category, among multitudes of other side projects with a historic fail rate of 80 percent. It's tempting to put this one in the same camp as Google Books or Google Scholar.

But, as repeatedly illustrated, Google products and services are often surprisingly related to the search business or are masks for some other, some grander, purpose. Google News, for example, Marissa Mayer estimates is worth $100 million despite its free status because of the number of searches it generates on the main search engine where ads are displayed.

That Google 411 thing? That was a front for voice recognition technology development. Callers got their info, Google got their voices to play around with. Indeed, Google's services are appearing less spontaneously altruistic and more mutually beneficial with every new launch.

Google Translation Center would be no exception to that new rule. In June, CEO Eric Schmidt said the goal was for Google to be able to translate 100 different languages, so this seems like a step in that direction.

But also, in the same way the Rosetta Stone was a key element in translating ancient Egyptian hieroglyphs, Google can use submitted human translations as comparison tools used for machine translation. Google's "payment" for use of its Translation Center, then, are the translated texts themselves.

Currently, machines lack the capacity to understand nuances in language, and therefore lack the ability to understand highly contextual, colloquial, or combinative search queries. Having human word strings and phrases to compare them to would help not only in the same language search capacity, but also in offering cross-language searching.

That helps Google in search, but it also fits in with the larger context of making all information available globally—except in China, of course.