Don't know of a roadmap, but I can say the 4.0 was the last version to have the RCA jack for the blower output, and the last version to have plugs on the bottom (rca blower and DC power). The 4.0 did not have a header for the servo, but you can wire directly to the board to make a servo work. The 4.0 was also the last version to have the old blower driver circuit that pulsed the ground instead of the +12V. It was a good board though, having less noise on the probe lines than the V4.1 board.
The 4.1 board moved the power to the side, removed the RCA jack for the blower in favor of a CAT5 jack that has the pinout for the blower and the servo. There was an initial attempt to integrate the thermocouple but the circuit was unrefined. The blower driver circuit was changed over to pulse the +12v and the blower now uses a regular ground. This board has a bit more noise on the probe lines, and it had an issue with false lid mode triggering if you do not do the (very simple) "white wire mod". It also had an issue where the trace for LED1 stopped just short of the mega, so you need add a wire to make LED1 work. All that said, still a good board that works very well, and I like the position of the CAT5 jack on the same side as the probes on this version.
The next iterations of the 4.1.X boards were very similar, just attempts to addressed some of the issues above, but I am not quite sure how they all worked out cause there is less out there and less feedback. V4.1.3 did have a rather serous flaw where one leg of the DC-DC converter was mistakenly shorted to ground, these boards will not work unless you cut that little bit of trace. It's not hard to do, but you better know to do it or you will end up scratching your head when you first power it up! These last couple 4.1.X iterations are fairly recent and people are still posting looking for feedback on how the 4.1.4 version is working out...
The 4.2.X that Bryan is working on now is kind of a complete rework. He moved all the power and possible interference generating items over to the left side of the board, and put all the sensor circuitry on the right side, and increased the ground plain in order to minimalize noise as much as possible. He refined the thermocouple circuit into the first viable working version, still as an add-on board in the first iteration I do believe. I don't think he was happy with the first draft, and through discussion and experimentation trying to eliminate as much noise on the probes as possible he ditched the onboard RF module slot (still available as an external module) in favor of RC filters on each probe, and it also gave him room to add the thermocouple hardware right onto the board (rather than in an optional module) and he also added the servo booster circuit to this board. As far as I know he is still waiting on the first set of that board to be delivered to him for testing... and has already mentioned doing some changes to the ground plain...