[voaie]

Given more and more custom fonts in the OSes/websites, maybe by using some new APIs, we don't need to specify everything in the Unicode standard. We can design a new font format or just a separate datafile, to store those locale-specific information. The Unicode code points then becomes parking slots for different fonts(with locale info to be registered). And We can use the standard/default datafile to keep the old info about the current unicode standard (say Unicode 8.0).

This is just my first thought. Seems that the job of ICU is transfered to the OS or web browser.

[voaie]

I think the Unicode standard should not limit the use of fonts. Instead, let the font or the additonal locale datafile tell us how to deal with those locale issues.