The Internet is not all English, although it’s easy to forget that when most of your browing history is to “.com” and “.net” sites. For years Greenlight Wireless has partnered with WorldLingo to provide dynamic translation of mobile content using the WorldLingo translation API. Recently we noticed new language pairs (specifically Arabic) and I had the opportunity to review a section of Skweezer that tries to guess the language of the page using nothing more than the top level domain. While not always accurate, the domain name a good hint of a page’s language when other standard language indicators are absent. For example, pages hosted in Mexico (.mx) are most likely in Spanish. Another example: while the official language of Azerbaijan (.az) is Azerbaijani, the closest language we have in our API toolbox is Russian, given that country’s history as a former republic of the USSR.
I’m sure some of this is not quite right, so I posted the data I have as a page in this blog. While I hope it doesn’t inflame national passions (should Canada be “fr” or “en”? Don’t ask a Québécois…) I hope this resource is useful and that improvements will make their way back to Skweezer.
