And then there are the small oddball countries who actually enjoy material in their original language. In Finland almost all tv-shows and movies are subbed instead of dubbed, and many people I know prefer (in addition to myself) english software instead of translated. One of the most horrible examples is Microsoft Excel which has actually translated the names of formula functions. Or GCC having translated error messages. Also having "," as a decimal separator instead of "." is sometimes a PITA. Helpdesking people over phone is much harder when you actually need to translate/guess menu items/button label texts etc on the fly from english to finnish. And then there is the problem that there is always some software in english, and its much nicer when the whole system is actually in the same language. And then there is the problem of just plain crappy translations, especially those made by some hobbyists.<p>English is lingua franca of the internet. Even if having own language is kinda nice, its just not that practical in globalized world.<p>nb. Localization is a lot more than just translations. Even if no translation is made, proper localization should be still made. Number and date formats are a common example. Another example I have heard is actually having different icons in different areas/cultures.
i did something similar to this once. i came to the same conclusion as shipley: trying to have other people edit your nibs is madness.<p>i chose to solve it in a different way, though. i kept all my localized strings in my own simple text file, and did away with nibs altogether, preferring to build all views and windows and menus in code at runtime. probably wouldn't work so hot for really complicated user interfaces, but it worked for me.
So by doing some translation effort, you can get perhaps +20% more users, but this should be compared with spending the same amount of effort on some other part of the program. Maybe some nice new feature would also have resulted in +20% users with less effort?