The primary reasons for I18N are shitty codepages and untraveled coders. We turned our app -- which needed a different build for each language only five years ago -- into an app which works in any language, even Hebrew and Arabic despite the difficulties and OS defects (Windoze and *nix) in supporting bi-di.

When the apps were first created, they were for an American market. They turned out to be good enough to go global. Then an I18N team was formed and we found al the problems and showed the programmers everything they had to look out for. Who knew that the Spaniards would get so touchy about a calendar showing Sunday as the first day of the week? That ain't anything that most programers concern themselves with. Except that the Microsoft calendar APIs themselves had no means to modify the display order.

Unicode is still a fractioned group with some still (wrongly) extolling the virtues of shitty UCS-2. Only recently was the right decision made to use a fourth character rather than surrogate pairs for additional display. How do you provide software that works everywhere when the infrastructure itself hasn't been built?

There are so many variants of Arabic that we've almost run out of letters to hang on the digraphs. There are so many geeks with their little hang-ups that we now have to deal with Cuneiform! Yes, Cuneiform was added in the Unicode 5 spec. While I rather doubt that any of our customers are going to use a writing method designed for sticks and clay tablets, we do have them increasing their reach to even more remote areas and we have to support all sorts of weird squiggly lines. We do that now. It's a lot less painful to support these days, all because of I18N.

In an ideal world the unwashed hippie coders would know things about the world beyond their cubicles and outside a 3-block radius of their flats but we don't pay 'em eough to ever actually get out there and see it.

[ Parent ]