UCE-8: fitting the world's living scripts into two bytes

A three-tier character encoding that reaches 8,704 characters in two bytes where UTF-8 needs three — with measured savings, the costs, and the reference implementation.