> I've found that sorting and shortening strings is useful. I like my regex working too.
Good. Let me know when a language supports unicode that does that.
Here's a hint: Where does ÞORN get sorted? After Z? Between T and U? After TH but before TI? Mix with TH? Sorted as Y? Mixed with P? Transliterated as TH? Transliterated as T?
By you thinking your language supports unicode, you write code that doesn't handle these cases. Your program will suddenly generate an error when faced with this data and your user will be unhappy.
On the other hand, by simply treating everything as bytes you know exactly how involved you are and need to be. You can avoid algorithms that depend on sorting characters (which seems to be locale-specific) and you can avoid algorithms that change case (which also seems to be locale-specific). That's becuase you're supposed to be avoiding these things anyway. Your language has made you lazy and stupid and the way out ISN'T to just be more careful- to just try harder. It's to stop worrying about this crap altogether.
If someone ever figures out how to do unicode right, or if this were an easy thing, I could possibly agree, but it isn't. Unicode is really fucking hard, and nobody has gotten it right.