Don't agree with you
CPU time spend to convert big endian / little endian just take a look at the cpu utilization: in normal everyday use it isn't more than some % rarely more than ten enough to convert. for cpu intensive programs like audio/video encoding they all have versions compiled for non legacy hardware.
I suggest:
- a program structure for legacy (68k) sofware that emulate a 68k or convert on the fly little / big endian as the fastest amiga processor is a 68040/60 running at max 75mhz I suppose that even a single processor at 1500Mhz would be many times faster than any real amiga
- a program structure for new software that use x86 specificity : the only issue for amiga is that it can live in any everyday PC.
Be realistic amiga has always suffered from it's hardware specifity (even if it was his main advantage too).
Amiga's miracle of the 80's is no more possible: nowadays standard hardware is many times faster that any more or less selfmade hardware.
His only chance is x86 there are many x86 compatible processors, take a look at cpu market these days:
- Power pc : doesnt exist anymore on desktop
- Sparc : incertain future (does sun will continue to develop him ?)
- Mips : the only computers that used them are retired since 2 years
- ARM : Could be a good choice but doesn't exist on desktop.
Think : anybody could make is own amiga from of the shelf hardware just imagine a 250$ amiga ! Who will buy a 600$ (or more) computer that is dead 14 years ago !
