First, in order for airport wi-fi to not-suck, you'll need a massive subnet with a TTL of no more than 30 minutes. Yes, I've been in airports where a /24 subnet was apparently just dandy...
Second, everyone who's in an airport seems to want to stream Netflix or something like that; I do hope that Netflix throws a peering widget their way, because the thousands of iPads in that airport will strain the pipe pretty efficiently.
Third, you're on a single collision domain, half-duplex, along with everyone else. 5GHz may help matters, but 2.4 will still be needed for compatibility, and if you're stuck on it, you'll probably get useful speed out of a dial-up optimized RDP session an an SSH window, but the only way regular web browsing is ever worth it is if you have some absurdly early flight (5AM takeoff or similar), at which point 'using my computer' plays second fiddle to the better activity: sleep.
Sorry, I've just never seen it worth it. I always load up my hard drive before I go, and I've never regretted it.
The airport: the worst place to be in the cloud.
It's an oversimplification to say that it's a single collision domain. Any decent enterprise wireless network uses overlapping access points that will automatically select and change channels based on automated detection of congestion and interference. Yes, there is always some level of frequency overlap, but that is easily addressed.
As far as address spacing goes, there's a number of scenarios in which a /24 can be just fine. Perhaps they are backhauling everything to a concentrator that performs NAT on a per-AP or per-zone basis, coupled with L3 roaming. A single massive subnet with maximum 30 minute session time is probably the worst possible design.
Enterprise mesh wireless has come a long way in the last five or ten years. Shopping malls and airports have wifi networks with multiple hundreds of APs in very high density so that they can gather wifi device location analytics via triangulation in addition to providing guest wireless access. They use it for determining foot traffic patterns and also storefront dwell time and conversion rates. For example, an airport will be interested in knowing where wireless devices are detected at a standstill in high densities, because they may be able to move things around or otherwise modify the physical environment to make it easier for folks to get to where they want to go. Interesting and also somewhat scary stuff.