It feels a bit weird, but dnsdir assigns reviewers for this alt-tld LC, too.  I took this as an opportunity to read the current text more carefully and also read more of the lengthy discussion history.

I'm very happy with how the rules for handling .alt turned out and I have nothing to suggest there, contrary to issues I see in texts for most special-use names.

Except a nit: I've never understood why mention null label in non-DNS wire format; I'd personally stick to a simple formulation, reducing the whole paragraph to just two lines.  The intention is simple - the draft does not define anything about how non-DNS protocols handle or represent anything.  Still, I don't mind the current eloquent formulation.

The major question is the high-level idea of alt-tld itself.  There have been very many discussions about that, spread over years, and last call would be quite late to significantly reconsider this anyway, unless something serious got suddenly discovered.  My takeaway is that - while I can't estimate chances of .alt ever getting really useful or popular (at least in the relevant contexts), the costs and risks of this RFC seem pretty low so I'm willing to give it a shot, perhaps partially as an attempt to deflect future discussions on DNS-like non-DNS names.

(A couple editorial comments went directly to the GitHub repo.)