As far as I'm aware, this is low-priority at the moment. However, I can see there is definite high demand for this feature, so in my opinion it should be kicked up the list of priorities right after bug-fixing (and just before 5B support)! With Famicompo 9 coming up, this would be quite a big asset to have handy.
I maintain that given the fact there is very high demand for it and that it is a perfectly acceptable feature for the NSF format, it deserves to be bumped up the priority list a bit. Maybe not before 5B as I said earlier, but not too far off it would be a good plan IMO.
I don't think yet another debate regarding multiple expansions will get us anywhere, so I'm just gonna say that the reason it's low priority is that it's a very special case that I don't really feel the desire for supporting. Adding support for YM2612 or SID as well would be just as legit in my opinion.
Saying it's a perfectly acceptable feature is also stretching it a bit, the NSF specification never mentions anything about using multiple expansion chips or the issues regarding it. There's also the fact that NSF was made for ripped game music and no games obviously used multiple expansions. The reason it still works is by accident rather than intentional.
Kidding aside, I probably said this before, but I would definately like to see multichip support. Of course, I agree with you in that it should be low priority, even if it's the very last thing you decide to impliment, but as long as it isn't forgotten about entirely. Bug fixing and support for other things should come first.
But regarding adding additional audio expansion chips to the NSF format and arsenal, there is already expansion header space available and it seems like the right thing to do if another commercially available device comes along.
For instance the YM2151 is now available for the Atari 7800 with the 7800XM module.
It's just that I prefer working on any possible improvements on 2A03 or other expansion chips before resorting to multichip support.
I don't dislike multichip tunes personally, it's just that I think it's close to being outside the scope of this tracker, and I can also imagine many other interesting combinations of sound chips (even though that wouldn't be supported by the NSF format natively) that could be added as well. Making a cart with a SID chip for example would be just as sensible as making a multichip cart, or using the powerpak to implement any kind of synth (within the limits of the FPGA).
