At this stage it's Speed-density / Alpha-N only and I don't have any real plans at this point for adding MAF support I'm afraid.
The easiest way if you really want to keep the MAF would be to use a hardware based frequency to voltage converter and such devices do exist. Otherwise, any reason not to convert to MAP?
They make the source available, but it's not under an open source license. You can't modify it, copy it, run it on your own hardware etc. They basically just make it available so that people can help find issues for their own products.
I looked closely at FPGAs and they have their pros, but they add a lot of complexity as well. Keep in mind that the original idea with this was to be able to load the entire system onto a cheap board and be able to test it on the bench without any extra hardware (Obviously not on an engine).
The other major reason not to use one is simply because these days I'm not sure they're really needed. The 'high speed' inputs to an ECU really aren't that fast, even at the highest end you're only looking at somewhere around 16Khz. With high speed MCUs (70+Mhz) available for a couple of dollars, the value in the FPGA is somewhat limited given the complexity of adding them.
There's a few. A big reason why Speeduino and RusEFI exist as open source projects is that Megasquirt, whilst they make the source available for viewing, is very much NOT under an open license. You can't use it on anything but their hardware, can't modify it, can't copy it etc. It's also quite hardware specific with big chunks of assembler in it.
Speeduino started as a way of getting a DIY ECU that was far cheaper than what Megasquirt offerred and it's currently about 1/3 the price. It's not on the same level for features at the moment, but I do new firmware releases every 2 months or so with more functionality etc.
The easiest way if you really want to keep the MAF would be to use a hardware based frequency to voltage converter and such devices do exist. Otherwise, any reason not to convert to MAP?