noyen1973 wrote:
From an emulator programmers perspective I would think it'd be easier to come up with your own coprocessor [...] than trying to reverse engineer someone else's and mimicing it.
No. You're always mimicking someone else's CPU. Fast CPUs aren't simple easily-and-accurately emulated devices. Unless you design something that
can't be built in hardware, in which case, again,
why bother? It's not a SNES game if you can't plug it into a physical SNES.
Quote:
That's all the coprocessors did but at least you could have it all under a single "roof", fully documented and simplified for coders to utilize and hobbyists to enjoy the coders' feats!
You're both overestimating the utility of a novel coprocessor and underestimating how restrictive the SNES is in the first place. The CX4 is already your math accelerator. The SA1 is already your faster main CPU. The MSU1 is already your streaming audio unit.
Quote:
The matching hardware may not precipitate but that outcome depends on the community.
The reproduction cost of software is cheap.
Hardware isn't. Hardware
must be the reference design. Otherwise there's no constraint on what one could justifiably call a SNES game.
Quote:
who's to say it might be another 10-20 years, or less the way technology is advancing, before we see a plugin cartridge that would of borrowed all of our ideas and made into a reality.
Me. I'll say it. It won't happen. Transistors won't get meaningfully smaller or cheaper. There are no more than a small handful of upcoming process shrinks, high-performance hardware won't get usefully faster. There's no pie-in-the-future-sky. Connecting what high-performance hardware we have to 5V systems will only get harder with time.
Design for manufacturability or go home.
—
byuu wrote:
A little known detail is that when I made MSU1, there was an existing similar project idea: a ZSNES Windows patch to connect it to Winamp to play MP3 background music there, and a Chrono Trigger hack to go along with it.
It also filled a niche that wasn't meaningfully handled. What niches still aren't meaningfully handled for which hardware could be built? Of those, which would meaningfully improve a preexisting game?