After a few of the recent discussions I started to wondered about the benefits and downsides to integrating a client side language translator such as .  Is the foreign market of players something game designers should try and incorporate?  Would the "all your base" syndrom make the attempts worthless?

Hmm. Well, first thought is 'any translation is better than none'. However, if it also translates it into the text of the culture, like Babblefish does when going from english to japanese, it can actually be worse than nothing. Not a lot of clients are going to be able to display kanji and even if they can, the person using the translation won't have the slightest hope of looking any of it up in a dictionary to see if the translation used words that seem to make sense.

A good translator imho should know when what it is about to send could be translated several ways and provide you with a list of options to use, not just assume the context from one sentence and hope it is right. Sadly, most users wouldn't want to spend the time to pick through possible words and find a better match, so there is never any option given to do so. This is the flaw in mechanical systems. The "all your base" syndrome is a symptom of human translators that use a dictionary and pick the first word they come across, without understanding plurals and other basic concepts. Machine translators usually get the right syntax, but pick the wrong words. Some are better than others, but all of them should allow for human intervention in word choice. Often, the description or associations a word has will make it very obvious when the wrong one is used. Without such consideration, the result is a disaster.

What's scary is that even megacorporations can't get proper translations right.

Pepsi once had their "Come alive with the Pepsi Generation" translated into "Pepsi will bring your ancestors back from the grave" in Chinese I believe.

You can only imagine how the various MUDding text would be translated.