I actually think in MSX 3 like a machine that I could use for several tasks I currently do with a PC: email, RSS reader, social networks, word processing, game playing, etc...
I mostly agree with Grauw logical step for MSX 3 would be:
- R800 or comaptible CPU
- V9978 (V9990 + V9958)
- MegaFlash SCC+ SD
- Ethernet UNIAPI interface
- Soundchips: PSG, OPLL, SCC, OPL4
- Maybe a new 16 bit expansion slot directly mapped to R800
Dreaming is free
Let's get into realistic dreams
The big irony is that the MSX is the most modular thing ever, but when talking the dream spec list, it goes like
every time you get a new box, you buy all the components again. That is inefficient.
The core deal of the whole story is a fast clocked z80. R800 is unavailable.
While all the rest exists, it is long known that MSX3 = 9958 + 9990.
A 20Mhz z80 has much higher 9958 performance than TurboR!
If there was a modular market, there would be 9958 cartridges.
And then TurboR users could plug that and too have more performance, SYNERGY!
It's a cool dream, but let's be reallistic: MSX3 is not a good deal. It should be a computer with no software, and for the comunity it would take a long time to provide a good collection of tools to make software.
If you really want to jump to another level in MSX development, you should consider to make games for MSX turbo R (with GFX9000 optionally) since it has too few games.
I agree with AxelStone. New software are targeted mostly for the barebones MSX1, and rarely for the MSX2. Absolutely no new software comes out for MSX2+, and one or another programmer supports the turbo mode of the Turbo-R (but none other of its features SCR>=10, PCM etc). Most of the programmers just deactivate the turbo instead of building proper timing routines.
For the extensions, the scenario is even worse (except maybe for the HDD/Flash interfaces):
- The OPL4 is the best supported extension, but still don't have widespread support
- The V9990 cartridges are the market for ages and is pretty much a paper weight
- Joymega had a very slow start, but seems to have gotten more popular recently
- The MSX-Audio BIOS was pretty much ignored by OPL4 cartridge makers.
- Ethernet interfaces are out for ages. Only Nestor has supported them. No 3rd party software ever used the nice UNAPI extensions he created.
More eccentric hardware performs even worse than the extensions:
- The Yamaha SFG-01 and SFG-05 has very few software support
- Franky only had support from the maker. No demos or games were ever made for it
- PlaySoniq sits in the same category as Franky. AFAIK, nobody ever used the built-in SID
- Musical Memory Mapper has no software for it
There's no shortage of new hardware for the MSX. But hardware without software is useless.
I agree with you and AxelStone, however…
- The SFG modules are not using a standard cartridge slot connector, so lack of support is unsurprising. But as you know, VGMPlay supports it, and it sounds really good
, so I’m glad Yamaha made it. Improving software support for the SFG was a big part of my motivation to develop VGMPlay.
- Franky and Playsoniq, to be fair, were made primarily to play Sega Master System games, and there is plenty of Sega software. MSX Synth supports the SID chip. Playsoniq’s giant memory expansion and SCC are generally useful (I’m really quite impressed by it).
- Musical Memory Mapper’s SN76489 sound chip will be supported by VGMPlay and Synthesix. Franky and Playsoniq’s, too. Tiny Yarou and gdx also made Sega SG-1000 game loaders which support it.
There's no shortage of new hardware for the MSX. But hardware without software is useless.
Hear hear.
If you really want to jump to another level in MSX development, you should consider to make games for MSX turbo R (with GFX9000 optionally) since it has too few games.
Imho it would be good if coders would target the Turbo-R as "a really fast MSX2+ with some extra's". And not rely on R800 CPU specifics.
Why? Going forward, real hardware MSXes will either be FPGA based like the 1chipMSX, or (homebrew?) machines that use some kind of Z80-compatible CPU. Maybe an 8/10/20 MHz Z80, maybe a Z180, maybe something else. But very, very, likely NOT a CPU that is 100% code compatible with the R800. Other Turbo-R specific extra's (hw timer, PCM functions, ...) are relatively simple, so should be easy to add to FPGA based systems, or perhaps add to an MSX2+ in the form of a cartridge.
I might change my mind if things on the FPGA side were different. But I know of at least 4 FPGA based Z80 implementations, and IIRC none of those has an R800 compatible mode. Which (MSX-platform-evolution wise) makes the Turbo-R pretty much a dead end.
As for new hardware extensions: why not code them as a patch to 1 or more popular emulators?
Much easier to do than hardware development. Easy to experiment with things that way, and (if patch gets accepted) you have an instant user base. If a feature gains traction & enough software uses it, then at some point do FPGA implementations or build real hardware. If a feature doesn't see use, then remove that code from the emulators.
More eccentric hardware performs even worse than the extensions:
- The Yamaha SFG-01 and SFG-05 has very few software support
- Franky only had support from the maker. No demos or games were ever made for it
- PlaySoniq sits in the same category as Franky. AFAIK, nobody ever used the built-in SID
Those I like most... ok, I'm an eccentric
Compared to V9990 or whatever rarely supported and owned chips, I think the whole Playsoniq thing is a very realistic option, with a part that you can use right away and another for potential. It's an all-in-one cartridge that gives a pragmatic power to the simplest MSX. Everything purely MSX is made easier, looks good, sounds good since you've plugged it in. The way it opens to the Sega Universe with an MSX heart, and vice-versa, is efficient and unique.
As for specific development that really takes advantage from its full power (MSX + SMS + SID + extras): I used to think it was a question of time, it's true that time has since past, but it doesn't matter much if it takes more (MSXsynth, Synthesix etc.).
It's not emulated yet, perhaps less attention is paid on for this reason? It seems SMS & MSX scenes have not joined efforts yet but mutual curiosity has raised, I think. Sure, there will always be people claming that marrying Sega, MSX and C64 is highly immoral and unnatural... let it be.
Personally, I think creating such hardware is a pragmatic and very good move for the standard to try and reach other standards: plenty of software to find on the way!...
And a more general issue about the hardware and the dynamic around : You never really know the popularity of the hardware before you start developping: did it sell well etc.
Sin
I see the MSX3 like that:
A box (power supply inside) with 1 cartridge port with I/O, 1 extension BUS with I/O, a slot to insert a android tablet (or a iPad) and other MSX standard connectors. The MSX turbo R is emulated by the tablet. The keyboard of the tablet is wireless.
I know it's a dream.
Raspberry Pi2 based to keep it in the spirit of MSX. Backwards compatibility for MSX1/2/2+/Turbo-R through emulation.
Don't get me wrong, I'm not bashing the extensions. It's the exact opposite: I think that we have plenty of awesome hardware with so little software for it, and my rhetoric was related to that.
When I saw VGMPlay was going to support the SFG-0x and SN76489 chip, I thought to myself: Cool! Finally!
@sinus
I agree that SMS and MSX scenes have a lot to gain if they join efforts. This close cousin of the MSX is also a nice machine.