Cosmopolitan Identifiers

The article mentions Cyrillic script, and offers an example of a translation of a CID from Cyrillic to the proposed Latin-based normal form.<p>This is tangential, but the KOI set of character encodings actually supports this out of the box, to an extent—if you remove the leftmost bit, you end up with an somewhat-readable Latin-script representation of your text. I thought this was brilliant when I first heard of it.

Lot of words for something that does already exist (C# for instance requires the first character of an identifier to be a letter or underscore but then most Unicode characters can be used), implemented in a more generic way (not limited to only three alphabets).