OCM and consequently, Zemmix Neo, SX1 and even the 2md gen OCM devices like SM-X and SX2, are far from perfect implementation of VDP
Technical difficulties aside, that seems to me an argument in favor of emulation, not against 
.
Not for me. I think it would be pretty silly to emulate the shortcomings of another emulator... Which also gets improved along the way.
I would be interested in using special features currently only available in MSX++ if more people had access to them, be it on MSX++ hardware or on an emulator. It's still an MSX which does things in an MSX way. I don't see why expanding the audience could be undesirable. It's kind of like with the ZX Spectrum Next. Having emulators available makes it easier supporting and making stuff for it. If you don't like it, nothing is lost. I don't see it as fragmenting the user base, but making it more attractive for people who, maybe, would be uninterested otherwise.
The MSXVR is something else since it's a RasPi running emulators. It doesn't interest me and I won't argue for it, but I'd say there is nothing you can do there that you couldn't do on an ordinary PC, and you don't need OpenMSX support to make stuff for it.
I think there are two different use cases:
1 - I want the super ultra setup like OCM, FM, SCC, Turbo z80, 4MB, etc... Add a 9990 and OPL4 to make it even nicer.... For that you can create a machine yourself OR use the beefed up machines already existing...
2 - I want to develop for OCM given feature so I would like to have it's switched IO and extras like vdp memory mapping: that would require someone very knowledgeable about those features to do it, or, the OpenMSX team to reverse engineer vhdl code... Former night be possible I think and Manuel may correct me if I'm wrong, but I see no reason why they wouldn't merge those new devices and allow someone to make a machine out of it... Later, I think it is unlikely at the moment, but the best bet for later would be to open a request in their github deitaling as much as possible and hope someone gets interested in picking it up
Everytine we talk about Next we should be careful, as it was heavily promoted and backed by Sinclair, there was a nice project before Next that was not as successful, not that I'm saying having an emulator doesn't help foster development for those machines, sure does, but Next is a special case on its own 
Well said, @ducasp. I'm trying not to be careless when mentioning the Next; although in theory the OCM was an official product released by ASCII, it never had the same level of support. It has been improved along the way, but despite the monumental and amazing effort by people like KdL (and others before him), a lot of people seem to despise the MSX++ or just outright hate it. The OCM's been out for 18 years already, for ASCII dinosaur's sake.
Some of MSX++'s features could be backported to MSX. What about a drop-in replacement for the VDP, F18A-style? This could be an incentive to further develop the VDP's compatibility, and bring new features to regular MSX models. It probably sounds like heresy to a lot of people, but really, why not? But I'm digressing.
I think a drop in replacement using OCM vdp has been done using tang nano, but I don't think it has the vram mapping as it requires quite a lot memory 
