Originally Posted by Galahad/FLT
As stated its going to be 68000 ASM using a Virtual Machine Emulator Core to run the Amiga code transparently.
Adding lots of flashy extras to compete with some of the bigger titles really isn't a problem.
For instance, transparent MP3 playback can be done with modifications to the emu core when certain "extra" custom registers are set on the Amiga side to tell the core to play an MP3, and other graphics modes with more colours are also easily done.
Mustve missed the bit where you stated you`d write in asm directly.
My point is that when adding to the emu and maintaining those additional features, then you could begin starting to think of it as a whole framework that could be turned into a (native) multiplatform one. Still a moot point as this aint helping much if your code is asm.
regarding graphics and colours, if you hit the chipset directly then you can run into troubles (setting colours in AGA is ugly enough already). if you dont (and use RTG or something similar) then you probably gonna write two versions/cores, one for running fast on Amiga-Hardware and one for the Emu?
Well, heres my thinking of a cross-plattform framework again - since you aleady have the overhead of doing "hardware-parts" twice.
Somewhere inbetween could be a compromise of only emulating the 68k since all the Chipset magic wont help you there and you get enough performance using libraries.
I dont want to sound overly negative, I wish you best of luck and would like a few new Miggy games. Just that starting something like this, using effectively multiple targets using assembler seems like asking for trouble.