yep. i'm playing with the same idea a couple of months. THE big problem: atari people won't buy a x86 base machine AND there is really no reasonable alternative. Maybe a powerPC board like the amigaone or pegasos uses, but that is far more expensive and slower than a x86. So in my opinion the AmigaOne is like a dead newborn baby.
So options are:
1. A 800(?)MHz of the shelf PPC based board (about 800€(?))
2. A 3GHz x86 of the shelf board (about €100 for an AMD based ATX solution up to €300 for a ITX (more compact) intel solution)
3. Coldfire board, does anyone make them?
4. Design a coldfireboard yourself (who has the equipment? Read about the logicananalyser czuba-tech bought - I can't afford equipment like that but i bet you need them bugtracking timingproblems)
So I would like a non existing option 5, a ready designed 3GHZ power4 board for less than €300 including processor (with emulation for old atari software an in due time a PPC native version of TOS/MINT) OR option 6 a MC68070 running at at least 1,5GHz. Other option swould work for some people but not for the masses. And without the masses we never see up to date software development continue. We need a decent browser (keep up the good work highwireteam, but where to get a java runtime environment for the atari that actually works?). Also I would like to see a good cubase clone for the atari that DOES work with obscure graphiccard solutions etc. etc. So my point: If we make something, it has to be something the masses are willing to buy. Of course atari people want to be different, but we do need a larger salesmarket than only the current atari users. (Let the flames begin)