Steve Rubel, Peter Hirshberg at Technorati and Richard Edelman have announced (only in English as far as I can tell) that Technorati and Edelman have teamed up
“to offer Technorati’s analytic tools in Chinese, French, German, Italian and Korean, starting with French in July and continuing into early 2007. That means not only will the user interface be translated into those languages, but the analytic tools themselves will be able to cluster blogs by language.”
A Technorati Japan beta (independent of the Edelman partnership) was recently launched, available to everyone, and “can understand what a user inputs in their native language and then give back related results.” As far as I can tell, in English, this means that you can’t search in these languages now and get results translated into these languages, but you will be able to as a result of this partnership.
Primitive Machine Translations Not Suitable for Business Use
What is not clear is how the translations will be done. The announcements sound as if machine translations will be used, augmented with local translations by humans. Given the extremely primitive state of currently available machine translation software, that’s a massive, and very expensive, undertaking. If it is what is intended, it would need to involve hundreds, if not thousands of local translators if the all-important nuance and localization of language are to be addressed.
Right now, machine translations are in a sorry state, generally not suitable for business use. And an incorrectly translated blog post could become a PR landmine in no time flat. Ethics Crisis, the blog I write for SRF Global Translations, gives multiple examples of the foibles of Google and BabelFish machine translations.
Edelman and Technorati are to be congratulated for taking steps toward globalization and recognizing that the world, indeed, is flat. It will be interesting to see how this plays out.
Related
Tom Braman has a great post on the problem with machine translations –
Swampy weed suggests whole state order recover open trust
To me it’s weird how they include CJK results in their search results. I don’t trust a machine translation at all.
The breaking down the language barriers theme isn’t new. Back in 1999/2000, We use to plug in Systran language translations modules into our chat rooms, message boards, and web polls so that we could have “instant” language translation.
These tools were used by over 1,000,000 webmasters around the world. It was a really neat feature, and was wonderfully received by the media, but they were not used very often.
One thing hasn’t changed: “Right now, machine translations are in a sorry state, generally not suitable for business use.”