Originally Posted by TCD
Or simply: Does it pay off to improve the performance on the Amiga or would people still buy the game without the extra effort put in?
If they're fun anyway, they will of course
Any game that's fun... well, you disregard the performance (which this thread is about surely). F.ex. Uno (being a Dan fan like me) played Gods to enjoy the graphics and then I... (oomph)...(wheeze)...suppose it works. But for me it was just too sluggish to enjoy.
Braben made only a few 16-bit games and spent years on them. They were ambitious with lots of effort in that direction. So his games are not quite like the 2D games with 10 bobs on screen and still struggling. It *is* simpler to write an assembler directive to swap scroll+bob routine between CPU and hardware mode, than for 3D because of (chipmem) polybuffers. It will be quite the smaller problem (for both bobs and polygons) if you plan for the hardware mode and port to CPU mode, but the main point is that it didn't happen (or any plans to use the hardware were aborted due to development cost reasons). Code one get one free! Best deal ever, except for Amiga users.
Same deal was available for PAL/NTSC sales, hence even arcade/action game ideas planned from the start to run only 25/30 fps.
So anyway, Braben is still a hero of mine for all sorts of reasons, not just being a great coder. It's just this frustration that so many games didn't take advantage of the beautifully conceived and balanced original chipset.
Speaking of Frontier, it would be a good "A1200 performance vs Falcon" game to pick actually. I'm not sure it would be fair to disallow 2MB extra ram for A1200 if it doesn't run on the 1MB Falcon, though.