I wonder what became of their other project, the Dragon.
It does seem rather odd that they would abandon it after getting so far along in development. The board they were showing at that one Amiga meet looked like the finished product and from the pictures appeared to be running Workbench and some other software.
There's pictures of it on their website. They had a prototype but ran into somekind of technical snag. Either it wasn't cost effective or they could'nt get it to work fast enough to be worth the effort.
At the time they claimed they were having difficulty obtaining a sufficient supply of the Coldfire processors, but that was BS. The most compelling argument I saw put forth was that the high overhead involved with the instruction trap approach to emulating the missing instructions made the price/performance ratio crappy compared to existing accelerators based on the 680X0 line. It doesn't make sense to me that they would get that close to production before figuring that out though. You would also think that they would just come out and say that if it were the case rather than just making excuses for a while and then going quiet without removing it from their site.
But from what I see of the progress on the Atari Coldfire project it should be doable.
The Atari Coldfire Project is interesting and it might be a good base for some sort of "Super Minimig", but I wouldn't get too excited yet. Right now they've got an Open Source version of TOS running on it along with a few test programs. Those are probably all compiled natively for Coldfire so aren't relying on the emulation traps. It will be a little while before we see what the performance is like for software compiled for older 680X0 CPUs.
