MSX dimensions

'MSX emulator' is the wrong term I think. It doesn't emulate an MSX, it emulates generic chips. And if those chips are programmed to act like an MSX, then so be it.

So the interesting debate is not strictly about the OCM but about when something is 'real'
- does something 'real' need to be a mass-produced unique chip that can do only one thing? (e.g. an MSX engine, a TMS, a Z80, a PSG etc.)
- can something called 'real' be housed in an FPGA chip, a chip which can be reprogrammed to be anything?

'MSX emulator' is the wrong term I think. It doesn't emulate an MSX, it emulates generic chips. And if those chips are programmed to act like an MSX, then so be it.

But how is that different from other (or: 'real' ) emulators?

So the interesting debate is not strictly about the OCM but about when something is 'real'
- does something 'real' need to be a mass-produced unique chip that can do only one thing? (e.g. an MSX engine, a TMS, a Z80, a PSG etc.)
- can something called 'real' be housed in an FPGA chip, a chip which can be reprogrammed to be anything?

retro devices are quite popular, and it's known there are currently many one-chip designs of old hardware around. Many of them not being reprogrammable, by the way.
For example, there are lots of so called "Famiclones", unlicensed one-chip devices that play Famicom (Japanese NES) carts.
Now, what if Nintendo would license such a device, and release it as a new NES console, with cartridge slot and all.... IMO it classifies as a genuine NES then (I've been able to play with one of those famiclones and they're rather close to the original). Accuracy plays a role, but I think to a certain degree (or you'd be able to say the MSX2 isn't an MSX either, because it doesn't run Galaxian and changes color palettes).
Is the fact the underlying chip (FGPA) is reprogrammable in the OCM's case a reason for it to suddenly not be MSX? If it was a not-reprogrammable one-chip design, would that make it more MSX?
And then there's the point of features... the OCM is quite the full-featured machine. Other devices, like the C64DTV, Atari Flashback 2 and Megadrive PlayTV (all based on 1 chip designs of the original HW) have severely butchered functionality (allthough a lot can be hacked back in by hobbyists, like cartridge slots). The Atari Flashback 2 (US release only) is perhaps the best example. It's apparantly very accurate, because, when a cartridge slot is added to it (which can be done relatively easily by enthusiasts) it plays original cartridges. It even has replicas of the original joysticks (that actually work on the old 2600 too). But it's lacking the cartridge slot by default. Would the lack of such a feature make it "not real"?

Not giving clear answers here because I don't quite know them myself... I find it interesting to discuss, though

I've never designed a chip using VHDL; but I think that it must be done taking care with the original chip design, I mean, all the quirks that the TMS VDP are due its design; clone it and you'll have the same behaviour.
AFAIK, the OCM doesn't have a 100% replica of the MSX chips; but I'm confident that the code will be updated...

It'll be the real MSX when it includes all the bugs/limitations of the original MSX. Same crashes, same shit -> same machine

But as far as it is an emulation, even a hardware emulation, it can only include "documented" features/bugs. All the "undocumented/unknown" issues won't be there, because it is not using the original chips with that genuine problems.

Does this make sense for you?

yes, and I understand that POV, but even the MSX1 and MSX2 differ a bit from each other... and the C64 for example had a revision in the SID chip, and the second is not exactly the same as the first.... but both are featured in genuine C64 computers..

Back on topic though, I think it's fair to not include it when talking MSX dimensions. I was also wondering if the SONY HB-700 is larger than a 8250.

Doesn't matter, they can't be as big as the National FS-5500.

Now THAT is a fat baby :D