What I mean by "accommodate the full complement of memory" is that it provides the full address and data bus to the outside world, where it can be utilized. When you have a development environment connected to a target system, you must ensure that you can address all the external hardware features of the target, yet not interfere with the operation of either the development software or the target hardware. You don't have to have all the memory in place, though the board to which I referred can, in fact, support all that memory, but you need to be able to access it all from the host processor.

I doubt very many target app's could benefit from as much memory as LINUX would require. I'm really just advocating that one use an ARM processor running LINUX as an ICE.

You're right in what you've said, but you should consider what I've suggested, i.e. slaving the target environment to the host-development environment running LINUX on the host processor, which has the same core as the target.

It's a quite specific way to "get the job done" but it's been proven by years and years of use, and it's just a way of leaving the PC's OS out of the process. Its actually fortuitous that the PC can, ultimately, though it's running a terminal utility, be used to enter new code into the system, though it's just as easy to do that from LINUX.

RE