The chipset uses the 0,1, or tri-state (floating) position to define the address bits. So as long as the receiver and transmitter matches, the packet goes thru.
IMHO, I don't think the "average" O-gauge guy will want to fuss with soldering the correct address pin of either the Tx or Rx DIP chips to the high or low or floating bus strips to match the address.
So comparing the 4-channel module RX (left) vs. the barebones RX (right), there is a substantial size difference. Obviously you need the 22uH inductor, bridge, cap, and some kind of regulator chip anyway. So if you use an SMT PIC chip (or whatever) rather than that monstrous DIP chip, I think you can build the board using the barebones RX for a similar size as the off-the-shelf 4-channel receiver...AND have a PIC to take care of any logic inversion, and other interface issues such as whether to interpret RC packets as toggle, latch, timed on/off, whatever. I stand ready to help define rolling-stock functionality from my perspective of those that could use remote-controllability.
As with the lighting board, I think there's great value in plug-and-play with minimal, if any, soldering. I think banks of addresses need to be done with either the 0.05" DIP style switches (vs. 0.1" spacing). I don't know how your lighting boards are doing but for modest incremental cost (since you always need the choke, bridge, capacitor, regulator) you can remotely turn off/on the lights, or dim the lights, or whatever. By setting the receiver board to the same address in each car, you could control/adjust the lights in all the passenger cars of a train at the same time. I think that would be quite the hat-trick!
As stated earlier, I stand ready to assist whether it be technical or moral support or the proverbial kick in the posterior if that's what's needed to move this along. It really is sooooo easy to imagine someone else doing this! 
