This was a good idea when it was part of the Application Bundle spec (or even 68k/ppc fat Mac applications), and it's still a good idea.
That's why I doubt it'll get much traction[0]. ELFs still don't even have an accepted embedded icon standard FFS.
Anyways, imagine what the world would be like if fat binaries were the norm and your OS guaranteed support for a "virtual architecture" that you could also compile to, that had the same interface to the OS as any native application. Then you could publish a binary containing native versions for all existing architectures and be assured that it would also work for all future ports of the OS to other architectures without the need for recompilation.
You could basically do this today with Linux, actually. Just pick one of QEMU's many targets to standardize on (and of course also standardize on a base platform of libraries with a stable ABI) and use it with the dynamic syscall translation.
[0] Turns out I was right, as this is an abandoned project. I wish I wasn't able to predict such things via pure pessimism.
That's why I doubt it'll get much traction[0]. ELFs still don't even have an accepted embedded icon standard FFS.
Anyways, imagine what the world would be like if fat binaries were the norm and your OS guaranteed support for a "virtual architecture" that you could also compile to, that had the same interface to the OS as any native application. Then you could publish a binary containing native versions for all existing architectures and be assured that it would also work for all future ports of the OS to other architectures without the need for recompilation.
You could basically do this today with Linux, actually. Just pick one of QEMU's many targets to standardize on (and of course also standardize on a base platform of libraries with a stable ABI) and use it with the dynamic syscall translation.
[0] Turns out I was right, as this is an abandoned project. I wish I wasn't able to predict such things via pure pessimism.