6502.org

It's a fair question, what level of fidelity is sought. Instruction-by-instruction is good enough for a great many purposes, like checking a Basic interpreter or a screen editor. To get games to do the right thing, you might need cycle-by-cycle. To get copy protection or demo code to work, you might need very precise access-by-access. (MAME's philosophy, for example, is to be extremely accurate, and never mind the performance cost, because the host platforms get faster as time goes on.)

I think you are conflating an emulator emulating software with an emulator emulating hardware. I (guardedly) agree with BDD. If you intend to create a hardware 6502 emulator it better be able to plug into the socket intended for a 6502 and take over it's job in every way including idiosyncrasies. If your ersatz 6502 does not have to interact with the real world, I guess you could call it whatever you like.

My definition:

An emulator of 'A' needs to do everything 'A' does for the purposes at hand. If that is just to run software written for 'A' precisely as 'A' (bugs and all) then that emulator could be in software too.

A simulator of 'A' only needs to look like its doing what 'A' does to the extent of the presentation. And yes, you could rig a simulator in hardware too, but software would be far more efficient.

This is the thing: we each have our own definitions, and our own preferences. There is no right answer, and no merit in seeking one. Anyone persistently promoting their preference as if it is objective fact is trolling. Deliberately trying to make trouble on this forum should be unacceptable.

Yet we should not discourage open discussion as long as it does not get contentious.

If I'm not mistaken, that's what forums are for.

Quite so. It's the strident insistence which is problematic, in my view. And then there's the question of going into unproductive tangents. (We know it's unproductive, here, because it happens over and over and the one summary topic we have is locked.)