If the 'addressee' is some human read blob that the shipping stuff and even postal carriers don't need to read, great. Heck it might be nice if everyone globally had their own UUID which could even get a lookup to 'current desired delivery address'.

Re: Unicode

Some sort of baseline protocol that just flattened everything out to what's really used and left behind anything remotely like Emoji ( E.G. Just 'single pen (monochrome imprint) compatible' Unicode ) might be useful enough.

I will likely forever absolutely loath how UCS-2 ended up baked into UTF-8 and the Unicode namespace generally _forever_ so that it gets kneecapped at an arbitrary low page size and yet still majorly complicates decoders that have to map around that inanity.

I think I'd want to review the 'Advanced Sub Station' bitstream format for inspiration of useful ways to extend 'font' operations. Any sort of 'color' or 'combining' operation would be relegated to _advanced_ / extended (not required) use cases and treated as a blob associated with either a single rune/character or a special meta character that denotes a sequence length. Among a bunch of other changes that would probably also include dropping UTF-8's 'resynchronize at next character' design in favor of some other mechanism that might include a simple checksum, and could change preferred encoding length at the next resynchronization.