I'm not an autonomous implementation designer; I am the servant of the
WG, which voted "yes" on "simple Posix", "TCP", and "UDP" and "no" on
"full Posix" and "full sockets". Feel free to propose an alternative
that satisfies these requirements.
As an example of sheer brilliance, I would point at your extremely nice proposal for how to deal with characters in a maybe-Unicode world, which WG1 agreed to. It's absolutely elegant, and does the Right Thing. The only thing I would even consider changing is that I'd like to drop the impossible char-upcase and char-downcase functions because I think it's more likely that they be used wrongly than anything else, but I understand why they're still there.
This is what kind of elegance you can achieve when you deeply understand Unicode, deeply understand implementations, and really can think clearly about a variety of different implementations. Note that while you leave some things as implementation defined, you are careful to do so it a way in which extremely useful stuff can be done even so.
The proposal you drafted for UDP, alas, seems to me rather like a casual one someone might have if they had a vague idea that Unicode existed, and said things like "string-upcase provides an implementation-dependent transformation of the string" and nothing else. I'll see if I can sketch out what something better might look like.
Thomas