TNI announces 萌ＳＸ

pfff 25 years. damn long time. party time where is the beer?

Sounds good! Is there already something to show?

Another marathonic endless project GuyveR800?, I wonder about some details....

By example, if the cpu WILL actually have a 32bit data bus, and the memory will read/write about 4 bytes per tick. I thinks it is the best way.

So the compatibility can be, if the MSB is just zero, it runs a z80 100% compatibile opcode. And 8 bit code will waste 3 bytes of each 4.

But, if the MSB <> 0, that one will be the z80 opcode, just using the same opcodes, but there is 3 bytes for parameters, now the actual 8bit parameters can be 16bit, and the 16bits parameter (nn) nn, can be 32bit (counting the two steps), being the first 16bits parameter for by example, a normal ld c, nn, can be ld csc, nnnn, where the last c same 8bit register.

It is a bit wasting, and not improvement in the ticks per instruction.

------------------------------------------------------------------------------------------

Other way
--------------------------------------------------------------------------------------------

shifting opcode, 8 bit databus, it also is a waste of time.

by example

pick a useless instruction, like ld b,b. This will broke the 100% of compatibility

but if (iirc ld b,b is $80).

$80 opcode params

This will extend the params to double sized registers. When working with that shift opcode, all flags must be connected to the MSB in operations of 16 and 32bits.

But as 8bit databus, it will double the cost in time. in both ways we can still work with the known opcodes, and it don't feels like it is not MSX.

So, which one this project will pick?

Like this site, TNI's does not render well on devices with small screens.

A new MSX would be so cool. I would support both in $ and in moral and hours. However I have doubts.
My reasons to doubt:
- designing a 32 bits processor is a lot of work;
- any MSX video chip is completely obsolete, you want HDMI and true color;
- audio (not mentioned) has progressed so much that even the best MSX audio card is obsolete;
- multitasking MSX-DOS3? Who cares about multitasking?;
- MSX-Basic was nice, but the way implemented it sucks. Better integrate a C(++) interpretor;
- backward compatibility: not being able to run Metal Gear 2 disqualifies for the name MSX;
- MSX probably can't be used because of copy rights;
- they talk about a research project. No results guaranteed;
- what is this machine supposed to do? (Internet? Text processing? Games?)

The architecture of MSX must be re-thought for a new purpose. For me, MSX is having fun with Legacy games.
So, I will follow developments

@Wolf, "a range of Z80 CPUs". Do you mean to give MSX co-processors for specific hardware? Do they need to be Z80's? or is any type of processor ok?

What I have been thinking about is, making a processor that does more "generic" more often used routines, like MMX was on Intel processors, for example a chip that can:
- copy a memory segment from ram to vram or v.v.
- do memory management
- Perhaps some assembly routines that now require lots of instructions (I thought Z80 division is decreasing until the remainder is smaller than the value)
- play music on psg/scc(+)/fm/audio/moon by simply uploading the audio files to the chip memory and start.

Like the amiga had co-processors for a lot of tasks (I/O, Audio, Graphics). The more intelligence you put in this "Z80" daughterboard, the better, because it keeps you Z80 free for the more "intelligent" work.

From a hardware perspective, it might be possible but some work on the motherboard will be required. Putting another chip on the bus, talking with the memories while the Z80 does its work. I imagine this to be a device with an I/O address, write some values to it and the work happens. This way: all old stuff works because the old stuff is still where it is expected.

Just letting my imagination do the work here

It is not my intention to criticise this project or other, don't get me wrong, but as usual many MSX projects from different groups with different specs and goals. There is no a clear idea of what MSX community wants/needs. It's not a problem of knowledge but organization.

That is because nobody has the same wishes for a new msx. Like many different topics here on mrc already mentioned, some think a turbo r + moonsound+gfx9000 is the ultimate msx, some want a device that looks like a pc out of the year 2000, while some dont want anything new at all.

Tbh from all the new MSX "follow ups", I think this one may be the most promising. If you look at the various specs, this one actually caters to what many ppl want:

- Faster basic (look at the recent renewed interest in Basic 'n (turbo/kun/bla))
- Combined v9958/v9990
- Faster CPU
- Multitasking is very nice, look at symbos popularity!

Well, I was just pointing out people like multitasking, not praising symbos itself (since I have never used it so I can't be the judge of that).

But the advantages of multitasking are pretty obvious imo. Having a multi-tasking DOS3 means you can run many already existing software at the same time and switching fast in between. For example, how about not having to exit your music editor just because you have to do some file management, or need to edit a sample. Or take a quick look at another sc5 picture while drawing without having to save/exit/reset. This is also a huge difference compared to symbos where you can only run software specifically written for it (instead of native MSX software).

Perhaps you need to experience it first. Like a mobile phone, or mobile internet. I didn't see the point in having those at first either

Btw imo multiple z80's would not be that great. Harder to code for, harder to design. I'd say one of the reasons people are still using msx in 2016 is its rather simplistic approach. But I am not an expert coder so maybe im wrong here.

multi tasking would be great for msx stuff. and how the make it possible i dont know. but as long as its easy in basic.

I mentioned SymbOS because that's a multi tasking OS, and it's obviously an office/desktop environment. That's not the same as running Nemesis 3 in thread 1, Moonblaster in thread 2 and Unknown Reality in thread 3. So, whether you'd need multitasking mainly depends on whether you need a Windows-style desktop for MSX, with dito applications. Personally I wouldn't ever use Office-type applications on an MSX anymore. In the 80's this was different, the MSX was our main/everything-computer.

You really feel your life is only complete when you can switch from a music editor to a file manager, a picture viewer or a sample editor? These examples really make no sense at all. :-) We can all boo at PC's, but these PC-things do exactly the things you mention. We don't need multi-tasking (which, as Fly points out, isn't all that easy for generic software), we do multi-computing!

For most people the point is money. All these new things cost money, and before it hits the mainstream, the gizmoboys bought the first few series, causing a massive pricedrop.

What I'd really like to see is an open (source) project. ATM, publicly, it's a news post, an e-mail address & IRC channel I guess. It seems they already made some decisions?
Not sure if they want to keep it behind closed doors (so to say), but I think/feel the most awesome thing to do is to make this a community thing/effort. (Of course there should be some lead(ing roles) as well.)

If they're really aiming for the next gen (of) MSX, I guess it would make sense to get a lot of people involved (anyone who cares essentially (and has some sense regarding the subject/matter I guess ))

Related, this is interesting perhaps?

and any extention on basic is cool!

Multitasking comes at a price.It often requires more memory (some primitive forms of memory management are required if you don't want applications to overwrite each others data) etc. Both these things: memory and speed are not abundantly available in MSX systems. When I look at Uzix, mostly the slowness of the system jumps out. With a 32 Bit system that will not be an issue though.

TNI probably has a plan and an idea and they go forward. Lets see what happens.

a better basic with split scroll. zoom rotate maybe split screen?

MSX basic is heavily limited by the limitations of the machine. Slow speed, limited memory, lousy debugging capabilities (" syntax error at line 130" -> list 130...)

A new basic can overcome these limitations, if you want to accept some incompatibilities, like MSX2 with 128KB+ memory only. For example: Basic is no longer integrated in a rom but its an executable on disk. 128 KB memory or more is available so a real editor allowing for good structured programming with decent functions and subs can be used. Code can be compiled before use. etc. When this basic compiler supports #include or linking libraries, any assembly routine can be used. It won't be a speed monster (even if you have a 7 Mhz machine) but your code will run way faster and perhaps: more people will dip there feet in the pool called programming.

What does my life's completeness have to do with anything? You asked for examples and I gave them. And I dont think my examples are nonsense. Multitasking is multitasking, the same for any system. I am puzzled why you think it is irrelevant for a MSX but great for a PC? If only because a PC can do it better, well we may as well throw away our msx because PC does 99,9999% of the things better anyway

Task switching may be something useful on MSX. This means: You can put the program you're running on hold, then open another program to do something else, to finally get back to what you're doing first. Maybe syn's example to Multitasking is the best example to Task switching:

The most interesting is, This could be done with Memory mapper and a little padronization.

I was thinking, if symbOS works under this new multitasking dos3, that would be pretty neat. You could run native msx software together with symbOS!

syn: you can't just run an OS on another OS without some kind of virtual machine idea...

SymbOS doesn't run under DOS.... it's a completely different OS on itself.

Does the CPU also have a protected mode? For multitasking this would be a cool feature.

Based on what are you saying this is vaporware? Over the years the creators have released a slow but steady stream of various kinds of software, so there is really no reason to believe this will not see the light of day....

Its a lot of effort but not impossible as some think. Besides, from what I gather, this is also a sort of consolidation of various other projects so a lot of work was already done in the past.

Let's not immediately pass a new initiative off as vapor, guys. It's a well-known fact that the TNI people are extremely knowledgable about MSX. Give some things a chance. We'll see, okay?

Are there any updates about this project? It seems quite interesting... I wonder how the progress is compared to for example the Vrobit.

Is there a public web (WWW) place that reports on / keeps track of this progress?

@ren: there is no active blog/status tracker or anything, your best bet is join the IRC chat at #msxdev on rizon (if you are using the webclient, try Firefox, Chrome seems to have problems connecting lately)

@tfh no problem :D

Is it dead, Jim?

Some information here: https://youtu.be/tx5uxSrE7iA

Yes, he also said himself he can use some help. Especially regarding the hardware part.

To be honest: to me this concept is the best so far to make a new MSX computer. The 100% compatibility is very much what I would also like to see in such a thing. Just a new affordable machine that can replace any real MSX and still run all software as if it was such a real MSX. And then add extra possibilities.

TNI is still working on it, as a matter of fact a lot was done last year/years.
There will be a website soon with more information regarding this project, so stay tuned

??

It is still being worked on.

Spacemoai (the creator) regularly posts updates on Twitter, such as these from earlier this year:

1. An indication of speed improvements https://twitter.com/spacemoai/status/1624666416333918209
2. Bugs being fixed https://twitter.com/spacemoai/status/1646176526340063232
3. And this one from last week about fixes to PUT SPRITE https://twitter.com/spacemoai/status/1655238796219232259

So its not "gone". It is just a massive project that takes a lot of time.

The more time he can spend on it the faster it goes, so if you support him on buymeacoffee he can spend more time on it.