Goolge’s statistical MT now allows for human (post) editing
Who ever looked at the best statistic based Machine Translation (and Chris from LinearB is adament that it is LinearB) will know that their MT process strongly suggests a post MT human editing step. We also know that Google is using statistical MT for some languages. So put those two together and you won’t be surprised at all that Google now allows a post translation editing stage for their “beta” languages - i.e. those that are done by statistical MT (including Chinese, Japanese, Korean, Arabic, and Russian). So what that means is that you can suggest a better translation if the MT is not accurate enough or, for example, mixed up nouns or stuff like that.
This is a fairly interesting development, I have to say, because the system acknowledges that there is no replacement for human interference when it comes to the complex task of translating.