If you find much of the “machine translation” versus “human translation” Internet debate tiresome and less than enlightening, then check out this refreshing, graphical, approach to comparing their respective merits from the New York Times ‘Technology’ section: Putting Google to the Test in Translation.
It takes a number of original texts and then compares the resulting translation from a human translator, Google Translate, Microsoft Bing Translator, and Yahoo Babel Fish using Systran. You can draw your own conclusions.
Of course, the usual vested interests will decry such ‘decontextualized’ tests as misleading and even dangerous. I don’t think so. I find these kind of articles more useful than some of the unedifying, outdated and mudslinging argument that all too often comes as as part of the debate on the Internet. For the ‘man in the street’ using simple, practical tests, graphically illustrated, gets the message out about the possibilities of machine translation (MT) to where it’s needed.
Exposing a broad audience to the opportunities of global knowledge sharing, elimination of information poverty, and generally increasing understanding between people is the most important part of the machine translation debate, often forgotten. Just look at the starting point when there is no MT at all.