After thousands of votes are cast testing machine translation engines on Gabble On, results seem to start taking shape. Here are some of the findings so far as reported by Gabble On.
As predicted in our blog Machine Translation engines put to the test!, it appears that survey takers prefer Google Translate’s results across the board. In a few languages like Arabic, Polish and Dutch, the preference is overwhelming with votes for Google doubling its nearest competitor!
However, once voters that have self defined their fluency in the source or target language as “limited” are removed, the contest became closer in some of the heavily trafficked languages. For instance, Bing Translator led in German, while Babelfish led in Chinese. Google however maintained a lead in Spanish, Japanese, and French.
Limited fluency voters chose Google over Bing by 2 to 1. They also chose Google over Yahoo Babelfish by 5 to 1.
For phrases below 50 characters, Google’s lead in Spanish, Japanese, and French disappeared, while Microsoft’s lead in German widened. Beyond 50 characters, Google’s relative performance seemed to improve across the board.
To eliminate any potential subjectivity toward a solution, on March 4th, Gabble On implemented changes to the survey hiding the brands and randomizing the positions of the results before voting. It will be interesting to see if brand bias played a big part in the results so far and how the new results will shape up.
If you have not done so yet, put the engines to the test and vote. Remember, one lucky winner will get the Apple iPAD electronic reader and eleven lucky winners will get Enabling Globalization: A Guide to Using Localization to Penetrate International Markets, the latest eBook in the industry on globalization and localization done the right way!