I think that clusterUK is on the right path. The necessary pieces are:
Standard PC hardware. As much as the specialized hardware was great for it's day, standard PCs have that now, and the only reason not to use the economies that the PC brings is to be different for difference sake, even if that means a lower quality product.
UAE integration. A large faction, understandably so, will only consider a new system an 'Amiga' if it is backward compatible to the A500. I don't know if I would go that far, but it is a requirement to get me to dedicate hardware to it.
Catweasal integration. Looks like this is on the brink. Being able to read and write actual Amiga disks IS part of being backward compatible, and is a key feature that makes the emulation look and feel native.
AROS for 68k. Ok, ClusterUK hasn't taken this on yet, but baring some bizarre rip in the space time continuum that makes all of the Kickstarts freely distributable, Kickstart for the 68k is going to have to be tackled eventually for backwards compatibility to be taken seriously. Today, it requires purchasing a separate product, or finding an illegal copy on the internet. While neither is difficult, it does prevent a freely distributable out of the box solution. Given how much trouble proprietary solutions on the Amiga have cause, this seems like a problem to me.
Getting UAE to be 100% compatible. Inevitable. If people use it, eventually all of the compatibility from WinUAE will be added to UAE.
Trademark. Either the Amiga name needs to be secured, or AROS needs to replace Amiga as the name of the platform, and the name Amiga needs to be seen by the community as a subset of AROS. Getting AROS for 68k working so that all of the MiniMigs could be shipped with AROS pre-installed would go a long way in that.
Once those things happen, I would say that AROS was THE new Amiga.
Doing all the same things to get C64 software running seamlessly on the same machine would take it to another level yet.
