MSX3 is Controversial?

To be honest I understand both sides... My take is that Nishi is kind of pissed off after the whole 9990 debacle and I don't blame him, being burnt because of a piece of hardware that barely sold and that there is so little software for it. It doesn't bode well for future software, too little for nowadays display, and why expand 9990 capacities to HD and waste resources when you can do much better and aligned to actual libraries/engines expectations? So he is avoiding like hell getting into conflict with the community (and my take is that his goal is to reach far beyond this MSX community and somehow get it to a PI / ESP status / crowd on the 0 and try to take a chance on the market with Turbo for general computing, and msx3 was kind of a love letter to the retro community that turned into a nightmare of sorts).

On the other hand community also had quite a few projects and news from Nishi that didn't turn into reality, so there is a lot of people that think this is just blurb and won't really be sold, others don't like the general direction towards making a device for today's needs with MSX name (because that means forget Zilog, discrete chips, embrace SOCs/FPGA/emulation for legacy), and let's be honest, every retro community is a bitch complaining about old days, old hardware feelings, wanting whatever to have a certain set of features and type of architecture, and even if you do that (MSX VR rings me a bell) still will find things to complain about (oh, it is a raspberry PI, my take, if it works great, my opinion, it doesn't matter the inside of it).

So, there's a natural friction between the guy that want to leverage a commercial success with an old brand that is out of the market so long and the elders that have their own vision of how it should have been and how it should be now...

Fact is, MSX is whatever the owner of the brand wants it to be, and whatever we dream off that is different, whether we work to make it reality or it is not likely to become a reality. Problem is that people like to complain and ask things, but you can count with your fingers, probably just one hand, how many really put an effort to create those things.

Not sure why Nishi isn't welcome in the Japanese MSX community, but what I've said above is pretty much how I see things between him and the non Japanese MSX community

I think ducasp is pretty close with his assumptions. A lot of people have a lot of different ideas about "MSX3". And even though most of us are in our 40's / 50's, we sometimes seem to be unable to have a mature conversation about such a subject, even though it's just a hobby.

It is impossible to please everyone. If you make an MSX3 aimed at older models, many people will say that it is an obsolete product. If you make a more modern MSX3, like an RPi, a lot of people will say it's not an MSX.

It's the same old story.

I agree with all these points.
Something that is great in MSX is that it's fun to develop with. You want to get something and you have all that hardware with their own rules all for you. It's an addictive challenge.

We tend to think that faster is always better, that multitasking is better, that programming a library is better than accessing the hardware directly, that high-level is better than low, etc. And yes, all this has its advantages, this is evident, but it doesn't mean that other alternatives should be ruled out. They can perfectly coexist.

I agree that having a new MSX which resembles much to a RPi running some Linux is not really interesting. However, if you offer me a machine with advanced hardware using a Z80, oh boy... I'd have much fun with it! Imagine that the BIOS (or a direct access to a port) would allow you to access several Gb of RAM only for you, without any multitasking or memory protection. And a large and direct framebuffer that I could use to render my graphics. Or perhaps some advanced VDP commands. Or a much more complete PSG. The key is that you interface with it as usual. If you want to make sound, you should interface with PSG registers, even if the hardware would be really different to the AY chip.

Keep backwards compatibility, add direct access, remove multitasking, add modern connectivity (USB, networking), and let everything be controlled by the BIOS and direct port access, and you'll have a machine you'd really like to develop with.

Well, this is what I'd love that the MSX3 is, but it's just my personal preference

I have not lost interest but I have decided it makes no sense for me to follow all the developments going left and right and every other direction. Considering the details lost in translation as well. I will wait for the end products (as there appear to be several) and then decide if this is worth to check out.