Not Lost In Translation

Google’s Language Translator is pretty amazing. You can give it a bit of text or a URL and it will spin it back to you in some other language. When you translate an entire web page it does a nice job of showing you the page pretty much unaltered except for the text changes. Most of this is possible because of the structure of web pages in general where all the pretty stuff is defined in HTML and CSS with the content stuffed into well-defined areas. The meat of it is yet another of the cool tools from the Google House of Smarts.

Take this poker site [nb: Deadlink removed] in Estonia for example. You can easily tell it is a Poker portal from the ads on the right and some of the links on the page. It does not have a lot of Flash and so is a good candidate for translation. Unless you lived over there or worked for a company in Espoo, you might not catch on right away that it’s likely from one of the Scandinavian countries. So you have to find a way to figure out the language.

You could make some guesses and eliminate many of the languages. It is certainly not from the Far East, Arabic, or Greek since those languages have distinctive character sets. It doesn’t have enough -ous and -ette endings to be French. Spanish and its relatives look nothing like this. It doesn’t have the tell-tale -ski signs of being Slavic.

That leaves just a few other languages that Google can handle. So, try each one of those to see if something that looks like English pops out. This is reasonable because Google Translate takes just a few seconds on each of these attempts. In a couple of minutes you have a page that shows the English text fitted into the original layout. You can be reasonably sure that the original is Finnish or some language close to Finnish.

They also do some other cool things with the page. You can hover over the translated text and it will pop up the original text. This is probably handy for those cases where you have some notion of the language and the Google Translator does it best but misses the mark or comes out with some odd grammar or technical terms. For instance, on this page while it knows that pokeri is ‘poker’ it does not know that nettipokeri means ‘online poker’ in Finnish. But even cooler, is that the pop up form offers you a chance to contribute to the translator’s improvement by suggesting a better translation for that phrase. Your translation suggestion is sent off to Google and added to the translation tool eventually [nb: It does not seem to work the same now, but you can still contribute translations].

So the next time you want to visit some strange poker site across the Gulf of Finland from Espoo, but would rather to do it in English instead of Finnish, you can use Google Translate to help out. If you are really feeling useful, you can also improve the translation for the next poor English speaker (at least somewhere in the future) by correcting some of the bad grammar or technical terms as when Google applies my change of nettipokeri -> online poker.

Of course, it would be nice if these sites allow players from the United States to play, but that’s just not in the game plan yet.

Imported from an old blog. Some links might be dead. Let me know if you find dead links.