I think it's a common misconception that only applications that are targetted at an international audience have to deal with the topics we usually think of as internationalization, such as non-ASCII character sets, handling time zones and international addresses correctly, and so forth.
But in this day and age, you can get most of these "international" data variations even from dealing with a strictly domestic audience. Most common word processors emit non-ASCII characters like directional quotes, and users are increasingly aware of how to make use of characters with dïacritics, symbols like ©, and so forth. Besides, if you're working on a web app that'll be going on the public internet, trust me when I say that you'll get all kinds of different data thrown at it from all over the world, whether you like it or not.
StickyMinds just posted my take on the subject as this week's weekly column: Bare Minimum i18n.