37 years since the Snatcher MSX2 version was released, that was ported from the PC88 one. Sense-san says he actually really likes the #MSX graphics which are wider and have better colors. What do you think?

@thara1129 @nishikazuhiko

So the idea of wrapping up the past by launching MSX2++ as the ultimate 8-bit ISA MSX, and then catching up the whole 16-bit era with one generation MSX2#, paves the way for MSX3. A 32-bit MSX for the 21st century, forward looking to 64-bit for the current era and perhaps beyond.

This means that MSX3 is still in the future. It's simply impossible, from any point of view, to develop MSX3 without any in-between steps and launch it as the successor to the 1990 MSXturboR. It would be difficult to set development goals, and it'd be hard to accept by MSX users.

#MSX has the aspect that "MSX3" has a mythical status. If you think about the expectations for MSX3 in the late 80s/early 90s, what will be MSX2++ now would've clearly been accepted as "MSX3" back then. Now that we find ourselves in the mid 20s, the ambition has evolved as well.

I thought this was pretty funny 😅

Back in 2022 a better YJK to RGB conversion was found. However, I found out that it still doesn't match a real #MSX machine. Using the below program, you'll see that the RGB border is slightly brighter than the YJK display area on a real MSX, but on an emulator they're identical.

あ、来週Xmasか？ 一年なんてあっという間だな

Since both YJK and RGB333 are converted to RGB555 in V9958, I wrote a program to visualise and count the number of RGB555 colors for various YJK to RGB conversion algorithms. New algorithm by @mdpc___ (19268 colors) Old algorithm (19268 colors) One of my algorithms (19617 colors)

Of course it's possible the 19268 number is wrong as well. Yamaha's manuals are infamous for containing grave errors. I should invest time to figure this out properly, shouldn't I? 😅

Back in 2022 I came up with a few different YJK to RGB conversion algorithms. I even found one that gives a similar result as a real MSX for the above test, but I didn't investigate further at the time. However, the alternative algorithms yield over 19600 RGB colors, not 19268...

Back in 2022 a better YJK to RGB conversion was found. However, I found out that it still doesn't match a real #MSX machine. Using the below program, you'll see that the RGB border is slightly brighter than the YJK display area on a real MSX, but on an emulator they're identical.

@nishikazuhiko @thara1129 MSXturboR# with R19K2terbo 🥺

@nishikazuhiko @thara1129 R1800 → R3600 → R7200 → R14K4 → R28K8 → R57K6 → RISDN

This naming change was decided by @nishikazuhiko and @thara1129 at MSX DEVCON 14 in Tokyo last weekend. I wonder how to pronounce it in English. In Dutch I'd say "MSX twee hekje", so maybe MSX two sharp, or MSX two hash? MSX two quad plus?

This naming scheme is similar to C → C++ → C#. On the one hand I feel this discriminates against MSXturboR, but on the other hand it makes a lot of sense because that was always a strange name. Also this way the crowd that shouts "MSXturboR is not MSX!" can continue doing it.

After letting go of the MSXturboR dual CPU architecture in MSX2++ already, it became a bit confusing for MSXturboR+ to exist in the hierarchy. This is why, consistent with the naming of MSX3, MSX3+, MSX3++ and MSX3# (4 plusses), the #MSX generation following MSX2++ will be MSX2#.

My first wishes: - V9990 I/O level compatibility - Sprite Linking with relative coordinates. There's a little known feature of MSX-BASIC for moving multi-layered sprites with 1 PUT SPRITE instruction. Adding relative coordinates, you could move and animate large objects easily.

HRA! opened a Discord server for allowing the #MSX community to make feature requests for V9978. It's the successor of the MSX2++ VDP V9968. The V9978 at the base level is V9968-compatible with added feature parity with V9990. You can help steer additional feature development!

It's amazing that @buppu2023 already added V9968 emulation to openMSX in a fork on github!