X8 vs X16 phase 3

[ZeroByte]

@The 8-Bit GuyI don't know how much work would be involved/ if Frank could quickly turn this around, but it seems to me that it would make TONS more sense to make phase 3 the new phase 1, and make the "real" computer become phase 2.

Take the X8, put in a beefier FPGA, and add the missing cores like 2151 and make it fully X16. Have 512K Hiram, etc. Then you can push out an easy-to-distribute product that doesn't split development efforts into supporting two parallel architectures. Use that time and income to finish ironing out the kinks in the proto 3 board/Kernal.

I say this without vitriol or resentment: Like it or not, the X8 WILL complicate development efforts for the ecosystem if it ever sees the light of day.

Even if it only takes a couple of hours' effort to port between X16 and X8, that's still a divide. I think that estimate is optimistic though. Petscii Robots benefits from having been designed to run on weaker systems like the PET.  Even with X16 enhancements, it's not the same porting this to X8 as having to strip or severely scale back features from programs originally designed with 128K VRAM + 2MB HIRAM in mind from the outset. Even for projects that CAN be easily ported, this still represents a cascading headache. If a project ever gets maintained after initial release, there are now 2 forks to maintain in parallel.

The biggest strength of the Commander X16 is the size of its community. Projects like this are 100% community-driven on the software side as far as the software ecosystem is concerned. Don't forget that this community is largely composed of Gen-Xers who have day jobs and family commitments. I.e.: development effort is largely done as a hobby in people's spare time.

Therefore, many if not most folks are likely going to find it too much of a hassle to maintain 2 forks of their projects. This means they're going to make stuff for one or the other system but not both. Thus both computers will have less software than they would otherwise.

That's maybe not such a huge deal for the intermediate/advanced users who're more Interested in making their own stuff anyway. The group this DOES hurt are the novices who would be mostly interested in getting their feet wet and in playing with programs made by others. Less stuff available means less reason to get either one.

Sorry for the lengthy post, but I felt it necessary to elaborate on this issue. Many have said "The X8 will fracture the community" but I haven't read anything explaining why that's likely true and what that means to the ecosystem.

Consider what the C256Phoenix would be if Stefanie had your reach, community, and fan base. Your system has more potential in my estimation than any similar project exactly BECAUSE you have the ability to make so many people aware of it. It truly has the potential to hit critical mass. This is why I chose to join the X16 fan club. I truly expect some day to be playing a high-quality game on it as a user, and enjoying a good selection of supporting apps like tool chains and API libraries as a developer. I don't foresee this on any other retro-inspired platform currently in existence or in the works.

I'm not saying the X8 by its very existence would wreck all of this, but there's a real risk that it could. If the Commander brand is destined to exist first as an all-in-FPGA system followed later by the premium all-retro board, much better that it's the same system I.e. the X16 "phase 3" and not a project-forking "Commander Gaiden"

[Snickers11001001]

This is a really interesting angle.   I like it! 

[VincentF]

?

This I believe is what we should consider instead of having a separate system.

I would rather preorder a phase 3 than a X8 ?

[BruceMcF]

The cost side would not change. Use a higher pin count, higher logic resources FPGA, it will cost more. Have it a two chip, FPGA and SRAM rather than a one chip FPGA only design, will cost more.

To me, the estimate that the CX16e would cost twice what the LX8 would cost sounds just about right. And the difference in market potential between being somewhere in the $25-$50 range and somewhere in the $50-$100 range (before keyboard and case) is a pretty dramatic one.

Now, if there is a higher logic resource FPGA that would support the CX16 Vera addressing, the "roving page" addressing system can be turned off (say, by directing it to the zero page, or if it is in the $8000-$9FFF range, directing it to the I/O page), and it had a soft YM2151 core, then it seems to me like it would be a wider range of things that would work out of the box and the porting of things that would be workable in a 64K/64K system would be substantially simplified.

 FPGA is totally after my time ... I don't know if there is a big brother to the Vera FPGA that has 256KB of SPRAM available, or enough block RAM to allow $0000-$9EFF to be implemented in block RAM rather than the SPRAM, but if there is, then a 104K/64K or 192K/64K version along the lines of able to use Vera access mode and including a soft OPM core would allow for quite a bit more cross platform compatible software. IOW, even 7 or 15 High RAM banks available would be a substantial step toward closer cross-platform compatibility over at most 2 (with bank 0 used by the Kernel). If the cost bump is less than 50%, for me that would be totally worth pursuing.