I am left to muse on why unlicensed, and even some licensed Western-made NES titles, are so badly implemented. I refuse to chalk it up merely to lol baddevs. The NES is harder to develop for than the home computer systems that many programmers started with. Many personal microcomputers had a design that mapped the video chip’s graphics into the CPU’s address space to make it easy to manipulate, but on the Famicom/NES, unless (I think) the mapper had some special magic in it, the PPU had its own memory that the CPU could only adjust during a certain time each frame.

How would programmers figure out this method of coding for the NES if they didn’t have access to any, or many, of Nintendo’s official documentation? Rare got started by reverse engineering the system themselves, but I have to expect not every developer had Rare’s technical acumen. I’m thinking especially about some of the games Mindscape made, especially their awful port of M.U.L.E., which is not only a glitchy mess but doesn’t even do a good job of implementing the original’s rules.

Joy Van most likely had to do what Rare did and figure out the details on their own, but as with most other companies who tried to do it that way weren’t nearly as good at it. I think it might be interesting, from a software archeology standpoint, to compare Rare’s early NES games to each other, and see if it’s possible to chart out their improvement in coding for the system. Perhaps Joy Van too, although I don’t know any of their work other than Metal Fighter, and even that only through Jeremy’s video.

rodneylives