Reading the "MSX Technical Data Book", there are some strange facts:

-The original MSX specification says "16 KB RAM minimum". There is a separate sheet that indicates that it should be "8 KB RAM minimum". Strange! If you take in count the system RAM, then you have less that 4 KB of free RAM!

-At the end of the book there are some machine code programs, and strangely enough, they are not in Z80 assembler syntax, but in Intel 8080's format. Was the original MSX designed to use this processor instead of the Zilog Z80?

-When it comes to decoding keyboard entry, it states that the BIOS calls should be used instead of direct hardware reads. "because the i8255 won't be used in future MSXs, that will have a wireless infrared keyboard". Amazing!

Just my two cents.

when one are designing an SO. / platform, one need to think in the future.

That's why they were using the Z80 in 1983, more than 5 years after its production?

What I never understood is the 3.58MHz clock... Isn't the Z80A capable of running on 4MHz? Is this a stability issue? Or what?

3.58mhz clock is the PAL color frenquency, they just not need a separated oscilator, simply can get it from the VDP.

Quote:

Thankfully, when Nishi announced the one chip MSX plans in 2001 he stated he aimed for an 'open' platform, encourating the users to not only passively use the system, but also to develop anything they want themselves.

That's great, but does it apply to hardware too? Will be people allowed to develop peripherals, or even clones of the one-chip MSX itself, without paying royalties? I know such a move would be difficult for a relatively small company like ASCII, but I'm affraid it isn't like there is much choice.

The desktop / notebook processor market has been long taken by Intel (besides not being ASCII's target, if I remember Nishi's statements right). The ARM standard already has a strong position on the handheld market, backed by big players like Intel and Compaq. I don't know how things are going for (mobile/smart)phones, but I would guess they're likely to move to ARM, as power requirements raise.

Under such a scenario, can a new licensed standard succeed? Possibly, but it certainly doesn't look easy. Again, I know how hard it would be for ASCII -- they're not a big mainframe company with no real interest on consumer eletronics, like IBM was when it freely licensed its standard -- but from where I stand, "open sourcing" the standard seems the best move to make it popular.

Then again, it could all be just me worrying about problems that might never arise. ^_^' I really wish ASCII and the new MSX the best luck, whatever way they choose to market it. It's just that a "free as in free speech" hardware standard, and that being MSX, would be so cool! Darn am I geek. -_-'

Quote:

That's why they were using the Z80 in 1983, more than 5 years after its production?

It doesn't matter where you are, little grasshopper, so much as where you are going to. ^_~

Using the Z80 made sense back then: it was still a decent processor, cheap enough, well tested and widely known -- so much it still gets used a lot, mostly on microcontroler architectures. If the MSX proved a success (like it did), it could always be swaped by an upgraded version (*). Of course, the less applications rely on the specifics of the hardware, the easier such changes get to be.

(*) I only wonder why the MSX consortium never tried to upgrade the Z80. Intel did it all the way from the 8080 to the Pentium 4; why couldn't MSX?

xperroni - Don't tell me you're unfamiliar with the great R800 processor used in the turboR computers. (pretty please? ). If 'R800' doesn't ring a bell, let us know, we'll tell you something about it, no worries...

About applying the open standard to hardware: yes and no. MSX Association (mind you: ASCII is no longer involved in MSX, Nishi packed his bags and raised the MSX Association) will license the one-chip-MSX, as they are the producers of it. Third party vendors will be able to license the chip for a fairly low price though. The CPU will most probably be ARM-based, while a part of the chip will be FPGA (and here comes the open hardware part), allowing software developers to adjust the hardware to fit their needs. I think, however, it's very unlikely that MSX Association would allow clones to appear on the market,...

I would like to see what an MSX could do if it had a 3GHz R800B, 256MB of RAM, a V9999 with 64MB's of VRAM and an OPL5. Assuming the whole thing had a 32bit bit architecture and perhaps a nice SCSI interface I guess it would kick frikkin' ass...

Hey sunrise! Any chance you can FPGA this for us all

Don't be too excited, Sonic, that never helps anyway

Quote:

xperroni - Don't tell me you're unfamiliar with the great R800 processor used in the turboR computers. (pretty please? ).

Oops, you're right. Yes, I know about the R800 and the turbo-R series -- the idea of building an architecture around two different processors still surprises me. But it was a somewhat late development, wasn't it?

Well, yeah... I guess we could have used it back in 1982, and if that were the case we would probably be posting on this forum with our MSX clones, but alas.... It's a kick-ass little CPU though.... Even the 'recently' released Z80 clones/improvements have trouble keeping up with it clock for clock...

Quote:

Oops, you're right. Yes, I know about the R800 and the turbo-R series -- the idea of building an architecture around two different processors still surprises me.

While the R800 is 99.9% Z80 compatible, its instruction timings are much faster than a normal Z80. A lot of software on MSX was timing sensitive, it needed each instruction to take an exact period of time in order for the program to work.

That's why a Z80 is still included in the turboR, because of compatibility reasons (and also because the MSX-Engine chip included a Z80 core, so it was basically free to add anyway.)

In order to prevent this from happening again, turboR programmers are supposed to use the turboR timer for delays and other timing critical things. This means in future MSX systems the processor used could vary in clockspeed or instruction timings without problems.

Unfortunately using the timer for timing is not practiced by most european developers (for instance MoonBlaster uses an extra PUSH BC/POP BC combination for delays), but all japanese software I've seen so far use the timer as specified by the MSX turboR standard.

So, building on the turboR standard, upgrading the R800 to 100MHz or 1GHz should be no problem
With such speeds, the Z80 can be completely removed and replaced by a software emulation.

Quote:

I would like to see what an MSX could do if it had a 3GHz R800B, 256MB of RAM, a V9999 with 64MB's of VRAM and an OPL5. Assuming the whole thing had a 32bit bit architecture and perhaps a nice SCSI interface I guess it would kick frikkin' ass...

Sure, but that I could do on an Intel box -- and without waiting months, or even years, for software built to explore its full power.

There is no point in competing for the desktop -- the future isn't there, anyway. As we move our operations to the Internet, looking for greater mobility, the need for thin, mobile, connectible clients will raise. That's where the MSX can make the difference, because of the extra it got: a developer community already used to many of the needs of this new market.

Is this future so far? I already keep many of my files on the Internet, on personal pages, in case I need them when I'm alway from my notebook. I don't bother downloading reference documents (like the Java API reference) anymore; browsing them on-line works just fine. And I recently gave up using an e-mail client: it's much better to keep messages on the server, and manage them through the web.

For these everyday, Internet-bound tasks, processor power isn't that important, when compared to band size, connectivity and standards compliance -- and those are the benefits a new MSX is better posed to bring. Let the server clusters do the heavy processing...

Quote:

HOW CAN YOU CALL ALL THIS FARCE AS REVIVAL !!!??? If there is no support at all for THE ORIGINAL MACHINE, MSX-1 then there is no revival at all.... MSX-1 was the thing which started it all... Revival means that lot of people are supporting and wanting to see something old fashion thingy, but how many MSX-1 revival games you have seen in last years, if Msx-dev´03 and Msx-dev´04 does not count...

Someone at MSX Association must have been reading this...