Public id discussions are so interesting if perhaps quite difficult to follow across threads, so perhaps this suggestion will have already been listed and dismissed, but here it goes.
the way I understand it there are two separate tasks with different purposes.
One part is authenticating an individual node (human users so far), on a self authenticating network. (I consider this an intrinsic quality of a network, node creation.)
The Second part makes those names useful, being able to find the app, the person, asset, the domain you want, consistently and reliably. This would be an emergent network property (nodes establishing relationships, communication protocols among each other)
Right now, from what im reading, it seems we are trying to solve the two problems with one solution based on the clearnet top down model of centralised DNS, with the self imposed limitations of that paradigm.
With that in mind , here is a path to explore:
For the requirement of the individual user authentication, a randomly generated keyID, non pronounceable (like a bitcoin address), is generated, with an optional field for name. Everyone can rush in to call themselves googles, apples, gods or turds, whatever they wish, etc. It can sound like a username, a real name, a registrar suffix, a post code an area number etc. But since anyone can call themselves anything , it wont make sense to search or trust any of these. They will hold no intrinsic value. but they do if they can be linked to the resource, (human, intellectual, etc) they claim to be or represent.
Design a way to group or map those keyIDs , perhaps using “proofs of …..” stake, human, resource, time* etc, creating webs of trust protocols. Maybe decorum is this...? And maybe use those protocols to create a decentralised verified address registrar, with datachain???? safe://theguardian.real …. being “real” the registrar name….? and real its suffix?
By separating the two functionalities we obtain infinite namespace, nullifying first mover advantage (making SAFE more inclusive to late adopters), and really make the network more flexible and adaptable.
I suppose the challenge then would be creating Web of trust protocols, not a small feat i would imagine.
This is really a half cooked idea, and I am quite the ignorant of technical details, but, just in case some part of this resonates with anyone here, there it is.
*maybe creating a external time keeping protocol by connecting atomic clocks to a public datachain!