Re: Unicode implementation

It might drag on this discussion, but it seems worth mentioning that variable-width encodings have 
been used for a long time, and string libraries for multi-byte encodings like Shift-JIS always use 
byte indexes. I am not aware of any library that used character indexes. Unicode has a lot of 
advantages over legacy encodings, but basic string operations still work the same way - you index 
string storage units, even if they are not bytes any more.

markus