I've been following the discussion here for probably well over ten years, now I'd like to contribute some to this discussion. Just a few questions, trying to get an idea of how all this works. So you take apart an Xbox and parts that are unique to the system can be broken down and examined and reconstructed digitally. So though I am aware this is half of the problem, I think, is the issue here the breaking down or the reconstructing digitally. In other words, if you had a blueprint of the console, would it be difficult to transition to the digital/ to emulate?
Then the question arises, okay, if you need a blueprint what limits are you facing just examing the hardware itself?
So okay, we have the hardware emulated so the second half of the problem is the operating systems that use the hardware? Clearly, we could just use the OS from the Xbox which I think one dev did. Is that what BIOS is? Are there, other than ethical and legal complications, major difficulties in lifting the OS from the system? Before you hang me, I'm discussing this purely from an academic standpoint. This is for educational purposes only.
Lastly, would it be practical to teach an AI to construct the language from the hardware, assuming the hardware emulated was indistinguishable? Stated otherwise, perhaps without emphasizing my complete lack of knowledge on the subject, how could AI learning tech be utilized to aid in the development of an emulator?
1/ Learning the hardware is hard as it uses undocumented chips, processors and such, reverse engineering requires expensive electronic tools and then trying to understand why the hardware works that way. It is a very difficult and painstsaking process that will take allot of time, imagining to do this on something like Sega Saturn.... Well that will be crazy.
Reverse engineering hardware is useful for devs on hardware they do not understand on specific hardware and why it can break games and when you fix those particular games it breaks others as in what they did for PS2 and it took the guy 6 months for one specific hardware.
2/ Software is difficult and requires reverse engineering now as just using the code made by the manufacturer will sue you pants of. Do not forget games use encryption.
From what I read long time ago original xbox uses lots and lots more instruction sets and libraries than windows so converting them while keeping the games run optimal at current hardware is difficult.
3/ AI is not an available solution, I mean it will help optimise the code but reverse engineering hardware and software and convert to current hardware and OS, AI do not miracously find a solution, it does so by making tons of mistakes until it gives a desired result, considering the hardware and softwware is different to anything else it will take allot processing and time so unless you got AI processing like used uni research, it will take a long time and even then will take doing it over and over as it wont give desired result until you figure out how to explain what the desired result in code.
Above all else all you have mentioned including AI requires one very important thing, man power with the time and knowledge. Not anyone can reverse engineer hardware and or software and have both of that knowledge and yet have that kind of knowledge plus work on AI, well that is rare.
Work is being done it just takes time and hope more people work on such projects but what is happenning of late is the devs are getting jobs and have little time to work on emulators, what is disheartining is the end user with toxic demands like it has been a year is the emulator done now, why isn't there android port and such... Now you can reach devs on discord and seeing that day after day the same questions when it will be done when it is done makes it harder and harder for devs to continue their work when this is a as this is their passion project and just let them do what they want to do, I mean with their experience and knowledge, they would know how to progess in their work.