With Google Reader's recent new feature, being able to translate all one's feeds into a language of their preference, using Google Translation Tools, it seems the language of the web is becoming less and less noticeable. However, reading the output of translated sites can often remind us of the value of knowing more than one language. The value I perceive is of course having expression for things which are not included in ones base language. French, the language of love has many shortcomings when describing technical situations, German, great for precise and complex descriptions but with a very different kind of romance, and for my self with asian languages, a word like rènào or 热闹 which has no direct English translation yet describes one of the key things I love about China.
When it comes down to it, Google translator sucks quite a lot, however languages do not make this task easy.
Where do we go from here?
I think there are 2 directions of interest.
- Write content with the intention of it being translated - For the time being we do not really know what this means I think most of us do not really know how Google Translator works, perhaps however Google could provide some guidelines on how to make things more translatable by their service. I think this could also include a feature which evaluates a given translation task and suggests words that are uncertain or sentience structures that may cause some problems.
- Create a systems of exceptions - If a writer tacks advantage of the specific strengths of a language it will become harder to translate, however those strengths are really important and can offer significant changes in meaning. To deal with this situation I think known exceptions could be added to a system like the Google Translation Tools. These exceptions would extend the ability of the translation system to firs state that there is some difference in meaning and secondly suggest words that may approximate the idea in the target language. The scheme of having user editable results (which google now has) is great but going on to use that data to create richer output would be really fabulous.
What do you think about this?