I'm a little late to the party, but I figured I might as well yell my thoughts into this void.
I have no problems with you dropping the case, especially if it's going to be a big issue. If things change in the future, I may be interested in buying the case separately if it is something where the finance issues can be resolved post-launch. As it is, however, not having a case isn't going to deter me from buying the product.
If you are going to half to hand-soldered the phase-1 boards, I'll probably buy the kit for myself. I don't really want to force someone else to spend that much time on just my board, and I'm willing to take the time to do it myself. (It's been a while since I tried soldering, but I think I should be able to handle it).
Between phase-2 and phase-3, the only
real advantage I ever saw with phase-2 was the one expansion slot, and I just don't foresee the expansion slot being all that big of a feature. I don't see ready-made boards or kits for X16 expansion cards being something that gets made, and the tinkerers who would want to mess with it would probably go for the phase-1 board anyways. With this in mind, I would say skipping phase-2 probably makes the most sense, but I could be wrong about how people feel about the expansion slots.
The X8 seems... interesting. Looking over it's features and limitations, a few things stand out to me. The 256-byte window into VRAM is neat, but poses some real cross-compatibility issues. It also says that it has "all of the same [VERA] registers", so I'd hope this includes the original VERA's data ports for X8/X16 compatibility. The use of USB for keyboards and controllers is actually something I sort of want to see on the X16. As someone that doesn't have any old PS2 keyboards anymore, I'm gonna have to rely on the one supplied with the phase-1 release, and as someone who never owned a SNES, nevertheless a controller for it, I'll have to go out of my way to get one to use with the X16. Using my existing USB keyboard and controller seems like it would be a simpler solution, honestly. Also, the bump from 8MHz to 12MHz seems... odd, and I would at least consider making it togglable back to 8MHz through a memory register for better X16 compatibility. Otherwise, it kind of seems like the X8 is both a better and worse machine than the X16, which just seems odd to me. I don't really care too much about the Yamaha chip being dropped. In fact, I kind of think it's something that should have maybe been dropped when the PSG was moved to the VERA. The VERA already provides more channels that I think I'd know what to do with, and I suspect the Yamaha might be underutilized on the X16, especially if the X8 does get released.
The RAM and VRAM shrinkage does stand out to me as being potentially problematic. The 64K RAM is probably fine. Nothing I've worked on so far required using the banked RAM, but that's also because most of what I've done has been pretty small-scale data and code wise. It's something I could see being limiting for larger projects, such as games or "productive" projects like text editors, music trackers, or assemblers. The 64K of VRAM
does seem somewhat problematic for me at least. Of the two projects I've shared here, only one would fit as it works now.
Noise X16 would work as-is, as it only requires 37.5K to store the video buffer (320x240@4bpp). If I wanted to use double-buffering for vsync, it wouldn't work on the X8, but I decided when releasing it that it wasn't worth implementing vsync for it.
AES X16 wouldn't work as-is because it requires 75K of VRAM for the video buffer (320x240@8bpp). The only real solution for this is to drop the vertical resolution to 200 to reduce the buffer to 62.5K, but the non-integer scaling just doesn't look very nice. I think I get why the memory is shrunk. I assume that the X8's combined RAM and VRAM is only 128K for the same reason that the VERA's VRAM is only 128K. I assume there just isn't any cost-efficient way of getting more memory onto it, so I get the limitation. I just don't know how much it will restrict developers.
I'm not entirely sure about releasing the X8. I
do think it will likely end up similarly to the C64/C128 situation, with a lot of software targeting the lower tier, but I worry it could be more complicated than that. I'd worry it might dilute not the
image of the X16 so much as dilute the specificity of the X16. To me, as someone who is too young to have really enjoyed retro computers in their prime, the X16 represents a modern-retro computer with an active community that will let me get as close to the experience I missed out on. Releasing a somewhat incompatible version along side the X16 will make the community less centralized, and would make it harder for me to get the experience I want out of buying the computer. It's certainly not a given, and I wouldn't say that you releasing the X8 would scare me away from buying the X16, but it's something I'd be concerned about if it does release.