The Amiga custom chips are still simple enough that their functionality can be easily understood and emulated in software or FGPAs (arguably, they're conceptually simpler than the C64 VIC-II and SID custom chips).
Every time I read something like this I remember back to the alt groups and people talking about the impossibility of emulation at all (if ever) of an amiga. What I find interesting about emulation is the amount of 'slop' that many programs can endure. Where the emulation is not quite right or even downright missing or wrong yet the program chugs along and just sort of works. Now some stuff needs that but it is kind of rare. Which is interesting.
I hear that claim, but it seems like we're still on the fringes. I understand we finally got to the point where you can build a C64 free of Commodore-exclusive parts only like a year ago. Is there a full suite of FPGA replacements for the Amiga chipset?
There's definitely some tangible appeal of building a robust '80s machine. Even with a quality software or FPGA emulator, the "feelies" are missing-- there's no opportunity to slide in an expansion card or manually fit it with a bunch of DIP RAM, or load stuff of of physical floppies.
Building a solder-it-yourself XT clone was a lot of fun for me, but I've sort of balked at the Amiga-flavoured projects in the space. It seems like they all start with "first get these five chips that really only can be harvested from a dead Amiga and cost a small fortune, so you'd better pray you don't put them in the socket backwards."
Maybe the middle ground would be using a FPGA to replace the custom chips, but the board is still designed with the right slots and sockets to fit an A2000 (or ATX) case and it still takes commodity parts like the 68000. Or maybe some 680x0 project targeting EmuTOS-- I'd expect it, as an open project, to be more adaptable to differing hardware than "must run exactly 1987 Amiga software with zany copy-protection and timing gimmicks".