There is a subtle difference in Apple's approach from that of Buffalo and D-Link, which also have introduced dual-band draft 11n routers. Both Buffalo and D-Link are resurrecting the "b/g for data, a for streaming" pitch that was used when the last attempt at reintroducing dual-band routers was made a few years ago. (That initiative was quickly aborted to start the 802.11n chase, unfortunately.)
If you look at the AppleTV and new Airport Extreme web pages, you won't see any sign of that pitch. In fact, the only thing indicating that the new stuff even supports the 5 GHz band is the small print that says the products are "compatible with 802.11a".
I haven't tested any of the dual-band draft 11n products yet, but the fact that the first three that have been introduced all have concurrent dual-band capability is a good thing. I've always maintained that consumer networking products are going to need 5 GHz capability to have a hope of providing reliable video streaming to the broadest customer base. And folks who have tried 11b/g WLANs and failed due to interference problems will be able to try again.
But hopefully, this time, products will have more intelligence and actually help the consumer have a successful WLAN experience by being more like cellphones. Even though your cell phone operates in two, three or even four bands, have you ever needed to even think about what band was being used, much less manually choose the operating band? Of course not! The phone and network take care of all that.
And so it should be for wireless LANs. The transition to 11n should bring with it the transition to "smart" multi-band (or as one chip manufacturer called it, "wideband") WLANs. Routers should default to automatic (and dynamic) band and channel selection, with manual selection moved to an "advanced" setup screen. Products should be smart enough to, without user interaction, set themselves up based on the RF enviroment, client mix and applications being used. Packet prioritization should be automatic and not rely on any client-side technology to get the job done.
If history holds true, I look for Apple to lead the way, but the reality remains to be seen. D-Link might have a shot, too, since it already uses Ubicom's Stream Engine technology, which I know can do automatic traffic prioritization on an application basis with no client-side software needed. But I'll withold any "attaboys" until I get products in to the lab and see what they can do.
Nevertheless, you should insist on dual-band equipment when you decide to make the move to 11n. Yes, initially this will cost you more. But you're going to pay more for 11n gear that really works anyway, so pay a little more and buy something that will give you the best shot at having a successful wireless experience. (Note that the low prices you see for first-generation draft 11n products are due to the fact that they're not selling and vendors need to clear inventory for the next round of gear.)
And if you're on a tight budget, then just wait awhile. As demand increases, new chipsets will appear that drive five-chip dual radio designs to three, then perhaps eventually one. And it won't be long before most dual-band wireless routers are well under $100 again (and hopefully, dual-band adapters as well).