I'm not in the future yet, but I was under the impression that with IPv6 you would actually know your own public address, because there is no need for NAT. Is this not correct?
NAT will still be possible - just not necessary for conserving address space.
But... imagine trying to get your call-center staff to ask the customer to read back their local IPv6 address so you can troubleshoot. That's going to be bad enough.
Then consider there may be proxies in the way, intentionally or otherwise, so you probably want to know the IP address the customer is hitting your product with, not their computer.
Most of us on HN are probably smart enough to find out our own IPs a half dozen ways - it's the 3rd party support that's tougher.
IPv6 addresses are statically allocated (well, in the IPv6 proponents' designs - in reality, we may end up with DHCPv6 and the like), but almost always provider-allocated (PA), not provider independent (PI); so you'll have to give it back to your ISP when you switch.
This is a good thing, by the way; PI space consumes memory on the backbone routers (they can't just send it to your ISP, after all - the whole point is that you can take it with you). Of course, PA space is just part of a larger allocation to your provider.
