"NAT, itself, however, evolved into a critical part of networking and much is built around it, not least the idea of a central device controlling access in and out of a network."
Firewalls and routing rules predate NAT by several years and both clearly involve the idea of a central device controlling access in and out of a network. I respectfully suggest that you present a fresh argument.
"The problem is that those pushing IPv6 view NAT purely as a work-around - a band-aid covering a problem of limited public IP addresses."
Perhaps they were around when the NAT RFC was published, and read it. I'm afraid that NAT *is* just a band-aid around limited public addresses.
Furthermore, not a lot of the coverage here is bothering to mention *why* the number of global routes has now passed 512K, so I'll let you into a secret. It is caused by people buying up small allocations of IPv4 one corner of the globe and using them in another. The address space has become horribly fragmented and the IPv4 internet is going down like a 99%-full hard disc using the FAT file system. And of course the reason everyone is still on IPv4 is because NAT has allowed them to punt this problem into the long grass for almost 2 decades. Well done NAT.