Google.com will speak native IPv4 for a very, very long time.
I think you should expect to see many, many sites that exist only on IPv4 for a while and are accessible via protocol translation services from IPv6-only sites, just as you're likely to see v4-to-v6 protocol translation for some new services that come online offered on IPv6.
But I've been through these debate wars before, and having been a central part of a handful of them and the various adoption discussions, there are lots of ways it can be done and they're all valid. Some people will hang onto IPv4 as long as they can, and they'll be reachable by both worlds (eventually). Some people will immediately go IPv6-only, and will be reachable by both worlds (eventually). Others will dual-stack, and be reachable by both worlds.
Not sure the bulletin software needs to support IPv6 (although it may do weird stuff with IP addresses). The web server / PHP combo that TMC runs can certainly speak IPv6 (assuming the Server: header is correct, 2.2.15/5.3.3). My guess is that it could be made to work.
- - - Updated - - -
I hate NAT, I really do and I love IPv6.
All I'm asking is for that AAAA-record for TMC
...and yet if you can already reach TMC, there's very little value in its AAAA record, other than as a demonstration. As someone who has regularly dealt with ARIN, RIPE, and APNIC to get address space (and the custodian of roughly a /12 of IPv4 address space) and one of the core organizations of the 6bone since the beginning, I do have a desire to see IPv6 adopted to relieve the mess. However, the real value is in the continued growth of the Internet of Everything, reaching rome geographies or verticals who must go IPv6. And even then, CGNAT's will work just fine for some time to come.
I could talk forever about this topic, but don't really have the time.