Jump to content

Change of product direction, good and bad news!


What should we do?  

374 members have voted

  1. 1. Should we release the Commander X8?

    • Yes, it should replace Phase-3. It's good enough.
    • Yes, but you should still offer a Phase-3 Commander X16 eventually too.
    • No, don't release the X8, stick with the original plan.
  2. 2. Should we still make a Phase-2 product?

    • Yes, Phase-2 is what I want
    • No, skip and go straight to Phase-3
  3. 3. For the X16 Phase-1, do you prefer a kit or a somewhat more expensive pre-assembled board?



Recommended Posts

The ecosystem for the X8 will be highly performant games and smaller demos.

This includes BASIC programs like Lunar Lander and Rogue-like games.  But I'll tell you, once I got CC65 working, I never looked back.  BASIC is not friendly for longer and more complex games.

Also, BASIC 2.0+ is no country for programs over 8K by people older than 12 years old.  Your brain has to be super spongy.

It also includes more cutting-edge assembly-coded games.  Maybe even some shooters resembling Wolf 3D.  

* * *

The ecosystem for the X16 will be larger, more capable games, and (potentially) richer small games.  They're more likely to be fuller featured, with a standard ADSR envelope library, fewer SD accesses (maybe?), blah blah.

Some fake examples:

  • Ultima X16 will have banked music and the "current local map".
  • The 7 Pirate Kingdoms of Gold  will have its entire map (1024 x 1024) banked, running in a vast simulation.
  • Traveller Trader  implements a generous swath of interstellar space with a high level of RPG detail.

* * *

Wild guessing.  Also assuming that all programmers cross over between two languages.

100% BASIC 2.0+
25% Structured BASIC ("ALGOL X16")
25% Forth
25% C
25% Assembly

 

So one in four developers for these platforms are working primarily in one of these, and secondarily everybody uses BASIC 2.0+ as the "batch mode" of the system.

 

Edited by rje
  • Like 2
Link to comment
Share on other sites

One application that would probably work on the X16 better than the X8 is GeckOS for the 6502.

 

The main reason I think it would work (much) better on the X16 is the benefits of all that nice banked RAM.

But, I could be wrong.

 

  • Like 1
Link to comment
Share on other sites

  

On 9/28/2021 at 2:49 PM, rje said:

It's hard to predict the breakdown of use with both machines.

That said, after I got CC65 working, I never looked back.  BASIC 2.0+ is no country for people older than 12 years old.  Your brain has to be super spongy.

100% BASIC 2.0+
25% Structured BASIC ("ALGOL")
25% Forth
25% C
25% Assembly

So one in four developers for these platforms are working primarily in one of these, and secondarily everybody uses BASIC 2.0+ as the "batch mode" of the system.

I totally agree regarding MS BASIC 2.0. It sucks lol... Even just the labels, functions, and subroutines of QBasic make programming so much easier to manage, let alone the nifty keywords that replace poke statements for various common tasks. Structured BASIC would be welcome. I've never used FORTH, but it sounds nifty. C would be great as long as the compiler is extremely efficient, given the memory limitations, and it's not bunged up with wacky object oriented syntax.

The main thrust though is that there are a huge amount of potentially interested normal people and a small amount of really big nerds. The nerds will totally buy an X16, regardless of cost and maybe even regardless of how much they would actually truly use it. The normal and even heavily nerd-leaning normal folks though, they're just not emotionally invested enough to care about something like the full sized X16 and when you add the high price tag on top of that general apathy towards the product, you've basically just created the perfect storm of disinterest... for the majority of people on Earth. And that's where the X8 would shine, it's cheap enough to interesting to average people, even if they're not entirely sure why they want it.

What would be an amazing product would be an X8 built into one those cheap portable game machine chassis. I bought one of those POWKIDDY Q90 devices for a whole $48 CAD and it plays NES to GBA games great. If the X8 came in that form factor and could be plugged into a USB keyboard and HDMI screen, people could program for and ON a neat portable gaming machine. That's cool!

  • Like 2
Link to comment
Share on other sites

Yes, I agree with you.  The reason I left the MEGA65 discussion lists is because I knew I wouldn't be able to justify buying one.

The X16 is more affordable, but it's the same issue: I pay a significant amount of money for something that's going to take up desk space that I don't have, and I'm not sure what I'd do with it really.

The X8, on the other hand:  I already want one.  It hits my price point so easily.  Worth it.  Regardless of whether or not I get the X16.

 

Edited by rje
  • Like 4
Link to comment
Share on other sites

On 9/28/2021 at 10:25 PM, rje said:

The X8, on the other hand:  I already want one.  It hits my price point so easily.  Worth it.  Regardless of whether or not I get the X16.

Well, I think this is what many of us thought the Commander X16 would be about in the end. It is just that an incompatible variant is not what we probably had in mind, but a single platform that would get there after an interation or three.

I wonder if the project might not be better off launching the X8 as the Commander X16 and focusing on that only, with possibly an X8 compatible larger memory version down the road, and forgetting about the hardware 6502 version completely. It would seem to me a higher volume X8 could quickly overwhelm any kit-based X16 ecosystem anyway.

  • Like 1
Link to comment
Share on other sites

I think folks who can get the X16 have already decided to get the X16.  The X8 won't change their minds -- they'll simply pick up one of those, too.

If anything, I think the X8 is a marketing tool -- sending out the X8 to evangelize would bring greater awareness of the X16. 

 

Link to comment
Share on other sites

On 9/28/2021 at 11:15 PM, rje said:

I think folks who can get the X16 have already decided to get the X16.  The X8 won't change their minds -- they'll simply pick up one of those, too.

Sure, but I do wonder which would be the development target focus? A slow trickle of more expensive kits being completed or an incompatible consumer-priced volume model being sold alongside it?

I think part of the appeal for the X16 is the idea of a new platform being born. If the volume focus is actually on a different, incompatible platform (the X8), I can see an effect on development focus and interest.

This is why I personally would love to see X8, if it comes, as a clear subset of X16 only and not a partially superior and incompatible platform compared to the X16. It would make the X16 more "worth it", in my view, if it was a "super X8" and not something separate.

Just my personal thoughts.

  • Like 2
Link to comment
Share on other sites

I posted this in another thread, but the FPGA that Frank has been using (ICE40UP5K-SG48ITR50) has been discontinued by Lattice. They still make plenty of versions of this FPGA in other packaging, but the physical designs for both the VERA and the X8 will probably need to be updated.

 

Link to comment
Share on other sites

On 9/28/2021 at 4:27 PM, Janne Sirén said:

Sure, but I do wonder which would be the development target focus? A slow trickle of more expensive kits being completed or an incompatible consumer-priced volume model being sold alongside it?

I think part of the appeal for the X16 is the idea of a new platform being born. If the volume focus is actually on a different, incompatible platform (the X8), I can see an effect on development focus and interest. ...

It's a hobbyist system in either case ... the "volume" model is not volume like a BananaPi, never mind a RPi.

Development is going to be hobbyist development. The people who want to develop for the X16 are going to do so if the system has a release date.

As far as whether it would be ideal for the X8 to be a strict subset of the X16 ... well, yeah, but if it's not a practical option, it's not a practical option.

  • Like 1
Link to comment
Share on other sites

On 9/29/2021 at 12:53 AM, BruceMcF said:

It's a hobbyist system in either case ... the "volume" model is not volume like a BananaPi, never mind a RPi.

Development is going to be hobbyist development. The people who want to develop for the X16 are going to do so if the system has a release date.

As far as whether it would be ideal for the X8 to be a strict subset of the X16 ... well, yeah, but if it's not a practical option, it's not a practical option.

Certainly my comparison of volume product is the likes of ZX Spectrum Next, also a hobby platform. It is an example of a single coherent target platform I have in mind. It has a couple of models but all are compatible. On the other end of the spectrum are the likes of C256 Foenix that has many low-volume variants and little in terms of clear target platfrom yet.

Personally I would prefer X16 project to resemble the Next in this regard: a single coherent target platform where possible model differences are details surrounding a shared, fully compatible core. On the Next there are even fully compatible clones of it, sanctioned, so it has a thriving software ecosystem.

Even hobby developers like to target an audience and this type of platform would seem to maximize that. Just my two cents on what I would prefer personally.

  • Like 4
Link to comment
Share on other sites

On 9/28/2021 at 5:23 PM, Janne Sirén said:

Personally I would prefer X16 project to resemble the Next in this regard: a single coherent target platform where possible model differences are details surrounding a shared, fully compatible core.

Agreed... and I think this is where we're at, if you define the core as KERNAL + system RAM + VERA, and of course the KERNAL upgrade has a call for copying memory to VRAM.

From 8BG's post, it seems likely that at least a thousand X16's can be produced.  Is that enough for an ecosystem?  I don't know.  8,000 ZX Nexts were sold, it seems, and I assume that it has a decent ecosystem.

 

Edited by rje
Link to comment
Share on other sites

On 9/28/2021 at 6:23 PM, Janne Sirén said:

Certainly my comparison of volume product is the likes of ZX Spectrum Next, also a hobby platform. It is an example of a single coherent target platform I have in mind. It has a couple of models but all are compatible. On the other end of the spectrum are the likes of C256 Foenix that has many low-volume variants and little in terms of clear target platfrom yet.

Personally I would prefer X16 project to resemble the Next in this regard: a single coherent target platform where possible model differences are details surrounding a shared, fully compatible core. On the Next there are even fully compatible clones of it, sanctioned, so it has a thriving software ecosystem.

Even hobby developers like to target an audience and this type of platform would seem to maximize that. Just my two cents on what I would prefer personally.

Yeah, the X16p and X16c would be that .the models around a single coherent target platform ... once a X16p crowdfund has launched successfully and then been shipped, and assuming that chip logistics issues have calmed down, a Quantity 1,000 and up X16c crowdfund campaign would be a natural next step. I expect with the X16p already in people's hands, a X16c with keyboard and case will fund at the minimum level to hit the keyboard minimum order without breaking much of a sweat.

Ideally, the X8 would be a pure subset of the X16, but the X16 is designed to run with distinct ASIC chips selected by chip select circuits made from glue logic, and the X8 is designed to fit into the FPGA that Vera was designed to run on, and they are a pretty different set of design constraints.

Going with the X8 makes the X16e redundant, unless the X16e is the board that is designed to be compatible with both the X16p and the X8 ... say, specify that the high bit of ALL ROMbank selections must be 0, and specify that X8 programs must never touch RAM locations $0000/$0001, and make the high bit of $0001 set the switch to select X8 mode in the X16e. The X16e will require an FPGA that is more capable than the Vera FPGA, so at worst that means that an extra step up is required than just to make an FPGA version of an X16p. Just as with the original three phase development path, that makes the X16e something that only happens if the success of the preceding phases justifies it.

In other words, that would make the X16e the One System to Rule Them All, One System to Find Them, One System to Bring Them All, and In The Darkness Bind Them.

 

Link to comment
Share on other sites

TL;DR: I'm new here and catching up on things, but wanted to share some preliminary thoughts in different areas in detail. Mostly about the cases, the X8, and the VERA.

Perifractic should make a case for the Raspberry Pis that's like the X16 case and has space for a decent fan like those RGB cooler tower ones. I definitely want a few. I've got one of these Argon AR2 cases for my RPi4B/8GB that has a noisy little tiny fan that doesn't cool much, and I'm just not so happy with the case design-wise for various reasons. Here's the thing. It hides your CSI ports. It hides your USB-C OTG port by going through the 5V and then through a non-OTG USB-C port so you lose data capability if you wanted to emulate a USB device through the USB-C port. So it then tempts you to leave your case unscrewed, and if you do, the way it's designed it wants very badly to bend your microSD if you just wiggled the bottom of the case just a tad, a little dab'll do ya, and if you do that, well, you could be looking at a potential fire hazard too because I found out one time on a RPi3B with an accidentally bent card it can heat up the unit very fast. I've got another NUC7i7BNH that has the little leaf blower fan that couldn't, that leaves the case hot, and it's noisy and might scare children and adults when it changes fan levels automatically at 3am, it sounds like a failing small jet engine. I took it out of the case, then found that the only alternative option for sale was a fan-less case, which is not something I care to do.

So I'm just simply saying, there's market potential here for better cases, and I rather liked the X16 case design.

I'm trying to like the X8, but don't know a whole lot about it, and how it differentiates itself from the C64 Mini. I've got one of those here too. If you start saying Raspberry Pi-sized FPGA, I start thinking IoT. The C64 Mini I found is not so good on the USB drivers. I finally got my Raspberry Pi Zero W to work with it as a USB boot protocol keyboard HID on the C64 Mini with some work, but the C64 Mini still doesn't work with my Arduino Nano 33 IoT with basically the same boot keyboard HID descriptor, it's a fiddly mess. I might get around that by changing the VID/PID, but I don't know what that's going to do yet in the Arduino IDE, as in screwing things up to reprogram it later. The more IoT-friendly the X8 gets the better. I start thinking IoT though and some words suddenly come to mind like UART/I2C/SPI, GPIO pins, Analog pins, PWM pins, 3.3V or 5V... I don't know what's all on the X8.

I can see that there are folks adamantly commenting in here that want things cut for cost to hit their ideal price point. The question then becomes, what is your ideal price point, and what can you do with it if you cut down cost too much, will you be able to do what you want to do with it? I'd like to know from those who want the X8 to be at a very low price point what that price point is and what they want to do with it. Make sure to test for feasibility, that's why that emulator is there. It maybe needs an update with an X8 mode for testing.

I was personally drawn to the X16 because of the YM2151+PSG and early on it sounded like it was going to have a 65C816, but I know today that's not the case. It all sounded kind of like a Sharp X68000 but in the U.S., which sounded way cool. I'm very much into some of the 16-bit era gaming console things like FM synthesis, graphical effects of the SNES like scaling and rotation, multi-layer parallax scrolling, and so on. The X16 was checking maybe not 100% of my check marks at first, but enough that it got my interest.

When I got into 65C02 ASM on the X16 I realized that on the X16 even though there was a 640x480x256 you couldn't do a full screen bitmap because that's 307,200 bytes. You could do that at 320x240x256 as that's 76,800 bytes, but now if you move the goalpost on VRAM to 65,536 bytes, now you suddenly can't do that again. You'd have to either scale down to 160x120x256, or use tiled mode afaik. Tile mode is fine if you have enough repeating sections of a screen. If you wanted to do text in a graphics mode, and had a 256-character font set, at 8x8x8bpp that's 16kb, and at 16x16x8bpp that's 64kb already. If the VRAM could go up to 76,800 bytes on the X8 and the RAM lowered to 54,272 bytes, or if that's somehow user configurable, which I would be surprised to see, that would be better in my opinion, because on the X8 you have faster file system I/O just from having microSD.

So, the 64kb VRAM on the X8 is a squeeze, and for some of the things I would want to do, that sounds like trying to squeeze a big orange through a small straw, and I want a bigger straw. Send me an X16 FPGA.

Should the X16 still be called the X16 since it's on a 65C02 as is? I remember those early 90s 16-bit console wars, I had a TG-16 back then and a neighbor of mine had a Genesis, and on the bus to school I'd have to hear, "well your TG-16 isn't a real 16-bit because it's an 8-bit processor with 2 16-bit graphics processors it doesn't have Toe Jam n' Earl in your face Blast Processing." Pepperidge Farm remembers. Later I got a Genesis and a SNES, the SNES was the best tech specs-wise and had some very good games. The TG-16 had a few good games too btw, and it at least had 8bpp color unlike the Genesis, but not that Mode 7 scaling and rotation from the SNES, or its SuperFX chip too from "muh-muh-muh jammen" Starfox, because according to Falco every English sentence is pronounced "muh-muh-muh jammen", anyway put that on a higher end VERA for a real X16, then sell it as a lower cost alternative to the Gameduino 3X Dazzler for breadboarding projects if that's feasible. Something that works with 65C02, 65C816, and 68000 chips for breadboarding projects perhaps, I'm not sure how feasible that is.

I think that both the through-hole kit and FPGA should be released. After all the Commodore had the Pet, the VIC-20, the C64, the Amiga...

Link to comment
Share on other sites

I wouldn't care much about a case. Design a board's form factor, bring that board to the market and if there are enough boards sold, cases will occur on the market on their own. The RaspPi also came without case and it didn't hurt it.

Regarding RaspPi, I think a lot of these are lying forgotten in some drawers, because people don't really know what to do with them. It's not powerful enough to use as a PC replacement (except from some nerds), programming is not easy to see some cool results for beginners, not much people have use cases where a server is needed.

What can an X16/X8 make better?

- it works out-of-the-box - no need to prepare an SD card first
- it comes with some primitive programming environment to do the first steps in programming, hence IMHO it might be more appealing as a school hardware
- programming it is much closer to the hardware than having Linux around it - similar to Arduino

What makes the Arduino successful?

- it can be easily programmed - simple plug and play to a standard PC
- good enough IDE software for the PC
- it is close to the hardware, e.g. not hard to connect certain other devices

What could make the X16/X8 superior to Arduino?

- you can play graphical games on a standard monitor (TV?)
- you can run different programs without the need to flash them from a PC
- you can program/learn programming on the device itself

  • Like 1
Link to comment
Share on other sites

On 9/29/2021 at 1:59 AM, rje said:

Agreed... and I think this is where we're at, if you define the core as KERNAL + system RAM + VERA, and of course the KERNAL upgrade has a call for copying memory to VRAM.

From 8BG's post, it seems likely that at least a thousand X16's can be produced.  Is that enough for an ecosystem?  I don't know.  8,000 ZX Nexts were sold, it seems, and I assume that it has a decent ecosystem.

The thing is, I believe "hitting hardware" is exactly the appeal of a platform like this. What differentiates it from coding, say, a RasPi. So personally I'd prefer hardware level compatibility between different platform models, not just through an abstraction layer like kernal, as much as possible.

But each their own of course, just adding my opinion to the poll. 🙂

Link to comment
Share on other sites

On 9/28/2021 at 7:49 PM, rje said:

The ecosystem for the X8 will be highly performant games and smaller demos.

This includes BASIC programs like Lunar Lander and Rogue-like games.  But I'll tell you, once I got CC65 working, I never looked back.  BASIC is not friendly for longer and more complex games.

Also, BASIC 2.0+ is no country for programs over 8K by people older than 12 years old.  Your brain has to be super spongy.

It also includes more cutting-edge assembly-coded games.  Maybe even some shooters resembling Wolf 3D. 

Don't underestimate the smaller BASIC/Assembler hybrid. There's a few machines like this, the Acorn machines, the Memotech MTX spring to mind. These allow you to embed Assembler in BASIC, and it can work quite well. The advantage, especially for the 6502 is significant.

  • Like 1
Link to comment
Share on other sites

On 9/29/2021 at 6:37 AM, Ju+Te said:

I wouldn't care much about a case. Design a board's form factor, bring that board to the market and if there are enough boards sold, cases will occur on the market on their own. The RaspPi also came without case and it didn't hurt it.

I think the sensible thing to do is to build it to a standard form factor. So the X16 could fit in a Micro ATX maybe, and the X8 in a Raspberry PI case, so people don't have to think about it if they don't want to. My requirements would be "something cheap and solid enough so it doesn't get lost or broken". If people want they can mount the board on a bit of wood or perspex, or acquire a custom made case.

One of the things that put me off Stefany's machine wasn't the machine itself or the $300 it costs ; it's that it's not a standard case layout (I think !) and the available one is half the price of the computer. And I just don't care about the case.

Link to comment
Share on other sites

On 9/28/2021 at 11:24 PM, john_e79 said:

TL;DR: I'm new here and catching up on things, but wanted to share some preliminary thoughts in different areas in detail. Mostly about the cases, the X8, and the VERA.

Perifractic should make a case for the Raspberry Pis that's like the X16 case and has space for a decent fan like those RGB cooler tower ones....

So I'm just simply saying, there's market potential here for better cases, and I rather liked the X16 case design.

That might be a different kind of thing than the partial case customization he was working on when he was a member of the team. Both the Micro ATX and the Mini ITX cases were customized front ends to existing available cases, where the customization would be done by the existing case maker so long as the quantity order was high enough (IIRC, 1,000 minimum order).

The X16 that we are seeing in the prototype boards is the Micro-ATX sized X16p, the one made as much as possible with through hole ASIC chips. Enough of the kind of people who would like to build their own X16p would also probably like their own case, plus the time required to build the X16p boards holding down the maximum number of built boards that might be offered in a crowdfund that hitting the required volume for the Micro-ATX case might be problematic. That is, at least, my speculation. In any event, the decision has been made to not go ahead with that.

Designing a bespoke case for a RPi would be either doing a new mold or else finding an existing case that could be modified, but unless the customization is 3D printed, the same volume of orders issue kicks in. And in any event, Perifractic would not be likely to go ahead with something like that without clearing it with the present design team. And really, unless/until the design team settles on their new plan for their development path, they wouldn't be in a position to lay out what parallel developments by Perifractic would be compatible/supportive.
 

Quote

I'm trying to like the X8, but don't know a whole lot about it, and how it differentiates itself from the C64 Mini. I've got one of those here too. If you start saying Raspberry Pi-sized FPGA, I start thinking IoT. The C64 Mini I found is not so good on the USB drivers. I finally got my Raspberry Pi Zero W to work with it as a USB boot protocol keyboard HID on the C64 Mini with some work, but the C64 Mini still doesn't work with my Arduino Nano 33 IoT with basically the same boot keyboard HID descriptor, it's a fiddly mess. I might get around that by changing the VID/PID, but I don't know what that's going to do yet in the Arduino IDE, as in screwing things up to reprogram it later. The more IoT-friendly the X8 gets the better. I start thinking IoT though and some words suddenly come to mind like UART/I2C/SPI, GPIO pins, Analog pins, PWM pins, 3.3V or 5V... I don't know what's all on the X8.

...

I think that both the through-hole kit and FPGA should be released. After all the Commodore had the Pet, the VIC-20, the C64, the Amiga...

 

How it differentiates itself from the C64 Mini is that it's an actual 8bit SOC system, designed on an FPGA platform, rather than being a software emulation of an 8bit system running on a 32bit SOC. It is a 64K RAM system with a small bootloader loading the Kernel & Basic from a serial flashROM, with the video system that is whatever X16 things fit into 64KB ... minus however much of the 64KB video RAM you use as extended memory.

As far as the IoT things on the X8, at present, basically none. There are two pins used for some form of debugger serial access to the X8, and maybe one I/O pin available (maybe not). My idea is that hopefully that debugger serial access pins can also be used as a two pin UART, and to use that extra pin to get multiple external selects on the SPI that is already used to reload the Basic/Kernel system from the flashROM and to access the SD card.

Someone else ... I think @Wavicle ... had the idea of overloading the debug pins with an I2C interface (since the FPGA that the X8 is built on has built in I2C modules as well as SPI modules).

But in any event, if it was used for IoT, it would be IoT programming pretty much from the bare metal in 6502 assembly code. That is not going to tick the boxes for 99%+ of people interested in playing around with IoT stuff, but that sounds like fun to me.

 

Quote

I can see that there are folks adamantly commenting in here that want things cut for cost to hit their ideal price point. The question then becomes, what is your ideal price point, and what can you do with it if you cut down cost too much, will you be able to do what you want to do with it? I'd like to know from those who want the X8 to be at a very low price point what that price point is and what they want to do with it. Make sure to test for feasibility, that's why that emulator is there. It maybe needs an update with an X8 mode for testing.

The point here is that there are several price points that is part of issue of whether to go with the X8. The X8 approach simply is cheaper than making a mostly FPGA based system that is compatible across the board with the X16p ... David's estimate is roughly half the cost.

The X8 already exists. Whatever price point it can be produced at is implied by an existing design. All of my hundreds/thousands of words spilled urging the opening up of the design to allow for a useful "hat" is just about tweaking an existing design. Indeed, none of it is about reducing cost, so much as about adding useful expansion capability at the smallest increase in cost.

The process of designing the X16e literally can't start until the X16p design is finalized, since its whole point would be to be to be the same to software as the X16p.

Remember, we aren't the ones designing any of this. This isn't a "crowd designed" project (indeed, crowds cannot design complex systems). We are just giving (in some cases quite extensive) feedback to the design team. So our opinions about hitting price points and which marketing approach are best are just that ... our opinions. In the end, it'll be the design team's decisions to make.

Edited by BruceMcF
  • Like 1
Link to comment
Share on other sites

On 9/29/2021 at 4:12 AM, Janne Sirén said:

The thing is, I believe "hitting hardware" is exactly the appeal of a platform like this. What differentiates it from coding, say, a RasPi. So personally I'd prefer hardware level compatibility between different platform models, not just through an abstraction layer like kernal, as much as possible.

But each their own of course, just adding my opinion to the poll. 🙂

Ah, I get your meaning -- that goes to the original intent of the three versions of the X16.  They're all X16s, even though they're built differently.

(As far as hitting hardware, well, I suppose the original X16e wouldn't hit any of the hardware, per se since all of it is mushed onto something like a single FPGA...)

 

I know that 8BG has a soft spot for Commodore, one I understand, so I suspect that it's not just about the hardware, and therefore an ABI for X8 compatibility at least seems a fun and useful idea!

Link to comment
Share on other sites

Quote

(As far as hitting hardware, well, I suppose the original X16e wouldn't hit any of the hardware, per se since all of it is mushed onto something like a single FPGA...)

Hitting the firmware?

Edited by BruceMcF
Link to comment
Share on other sites

On 9/29/2021 at 7:39 AM, BruceMcF said:

That might be a different kind of thing than the partial case customization he was working on when he was a member of the team. Both the Micro ATX and the Mini ITX cases were customized front ends to existing available cases, where the customization would be done by the existing case maker so long as the quantity order was high enough (IIRC, 1,000 minimum order).

The X16 that we are seeing in the prototype boards is the Micro-ATX sized X16p, the one made as much as possible with through hole ASIC chips. Enough of the kind of people who would like to build their own X16p would also probably like their own case, plus the time required to build the X16p boards holding down the maximum number of built boards that might be offered in a crowdfund that hitting the required volume for the Micro-ATX case might be problematic. That is, at least, my speculation. In any event, the decision has been made to not go ahead with that.

Designing a bespoke case for a RPi would be either doing a new mold or else finding an existing case that could be modified, but unless the customization is 3D printed, the same volume of orders issue kicks in. And in any event, Perifractic would not be likely to go ahead with something like that without clearing it with the present design team. And really, unless/until the design team settles on their new plan for their development path, they wouldn't be in a position to lay out what parallel developments by Perifractic would be compatible/supportive.
 

I'm talking a case that's totally separate of anything to do with the X16, totally new mold, just in somewhat of a similar kind of style but a new and different design. I'm only mentioning that because if I go to a retailer like my local Microcenter and look at the selection of Raspberry Pi boards they have, it's not all that impressive of a selection when it comes to cases. They all seem to have a tendency to check a few boxes and then leave out some important things.

So if the X16 prototype is a microATX form factor can't you just pick up any old regular microATX case from a Microcenter or Amazon or wherever then? If it's a kit that comes as a maker board kind of like the PE6502 probably anyone who solders parts together is going to just want to buy their own microATX case anyway is my guess. If that's all true, it just doesn't seem like that big of a deal then.

Quote

How it differentiates itself from the C64 Mini is that it's an actual 8bit SOC system, designed on an FPGA platform, rather than being a software emulation of an 8bit system running on a 32bit SOC. It is a 64K RAM system with a small bootloader loading the Kernel & Basic from a serial flashROM, with the video system that is whatever X16 things fit into 64KB ... minus however much of the 64KB video RAM you use as extended memory.

As far as the IoT things on the X8, at present, basically none. There are two pins used for some form of debugger serial access to the X8, and maybe one I/O pin available (maybe not). My idea is that hopefully that debugger serial access pins can also be used as a two pin UART, and to use that extra pin to get multiple external selects on the SPI that is already used to reload the Basic/Kernel system from the flashROM and to access the SD card.

Someone else ... I think @Wavicle ... had the idea of overloading the debug pins with an I2C interface (since the FPGA that the X8 is built on has built in I2C modules as well as SPI modules).

But in any event, if it was used for IoT, it would be IoT programming pretty much from the bare metal in 6502 assembly code. That is not going to tick the boxes for 99%+ of people interested in playing around with IoT stuff, but that sounds like fun to me.

Maybe not 99%, but I'm in that 1%, and I think that number is going to grow because you can get away with doing BASIC at first but eventually you might want to do some more performant things in ASM. Totally agree on reusing the debugger as a UART or I2C.
 

Quote

The point here is that there are several price points that is part of issue of whether to go with the X8. The X8 approach simply is cheaper than making a mostly FPGA based system that is compatible across the board with the X16p ... David's estimate is roughly half the cost.

The X8 already exists. Whatever price point it can be produced at is implied by an existing design. All of my hundreds/thousands of words spilled urging the opening up of the design to allow for a useful "hat" is just about tweaking an existing design. Indeed, none of it is about reducing cost, so much as about adding useful expansion capability at the smallest increase in cost.

The process of designing the X16e literally can't start until the X16p design is finalized, since its whole point would be to be to be the same to software as the X16p.

Remember, we aren't the ones designing any of this. This isn't a "crowd designed" project (indeed, crowds cannot design complex systems). We are just giving (in some cases quite extensive) feedback to the design team. So our opinions about hitting price points and which marketing approach are best are just that ... our opinions. In the end, it'll be the design team's decisions to make.

Here's a question, does the x16e have to be the same software-wise as the X16p? You mentioned in an earlier comment the need to move to a better board for the X16e than the existing VERA. So for those of us in the crowd of wanting things like more VRAM, maybe have a X16p compatibility mode, but also allow us to have X16e specific games and apps utilizing whatever features come from a VERA 2.0 you might call it.

If you had enough VRAM for a 640x480x256 bitmap, so 307,200 bytes, you would have enough for 4 screens of 320x240x256 hardware scrolling, maybe even more VRAM for more layers for parallax scrolling and/or more sprites. Sounds like it would better complement other features like the YM2151 and SNES controller ports, considering all of what the SNES PPU offered, which is even more than that with features like scaling and rotation.

Then you would basically have the lower end (C64-ish X8), the middle (TG-16-ish X16p?), and the high end (Amiga/SNES/X68000-ish X16e?). That seems like a winner to me.

  • Like 1
Link to comment
Share on other sites

On 9/29/2021 at 3:47 PM, john_e79 said:

Here's a question, does the x16e have to be the same software-wise as the X16p? You mentioned in an earlier comment the need to move to a better board for the X16e than the existing VERA. So for those of us in the crowd of wanting things like more VRAM, maybe have a X16p compatibility mode, but also allow us to have X16e specific games and apps utilizing whatever features come from a VERA 2.0 you might call it. ...

I don't know about "have to" ... I'm not in the design team after all .. but the goal for the X16e was that it would be software compatible.

Now, whether it could ALSO have upgrades that become possible when the video chip and CPU are literally built in the same FPGA ... well, I suppose it COULD. The "e" might stand for "enhanced". Indeed, it could have a choice of either the X16 or the X8 or both methods of accessing video memory, and there are bits available in various places of the Vera specification that would allow for using more than 128KB of Video RAM, so you might have "fast" Video RAM in the first 64K accessible through the page access as well as "extended" video RAM available in the X16 data port.

After all, whether or not the design team wants to fragment the hardware platform like that is for them to decide.

I am ambivalent about an "X16+" approach to the X16e, but I kind of like the idea of going ahead with the first X16p and the X8 in parallel, then the X16c, and then if the whole project is going well enough to keep it going, go ahead to an X16e that would also feature an X8 compatibility mode, so it would be One Board to Rule Them All, One Board to Bind Them, One Board To Bring Them All, and In The Darkness Bind Them.

  • Like 1
Link to comment
Share on other sites

"Don't know if anyone reads here *and* the forums but a chap posted this on there today :

"you mentioned that you guys already have an X16 implementation based on FPGA - wouldn’t it make sense at this point to go for that?"

and David responded with:

"Not exactly. We had the X8, but decided it wasn't quite compatible enough. So I think Frank was going to work on the full X16 implementation."

Has a decision been made I wonder ?

  • Like 3
Link to comment
Share on other sites

I'm new to the boards, but I'll throw my 2cents in.
Like others, I'm looking for something modern but as close to a system with discrete parts. The X16 is a different beast than what the Mega65 people are doing.

I fully appreciate the work involved in getting this thing made and shipped to people. It makes you appreciate the complexity of manufacturing a complex electronic device. Commodore made millions of C64s and had to juggle many of the same issues (though many different today) with the logistics and labor of a system that is not an SBC with an FPGA doing most of the work.

You can't expect to make a fully created and mass-manufactured discrete computer without..well..what a company has which is offices, warehouses, employees. I'm not saying it's impossible to do it with a team of distributed people who are dedicated, it's just going to be HARD.


I like the Mega65, but the reasons I'd want that system are TOTALLY different for the reasons I'd like the X16. The X16 is what I want so I can learn and tinker. I'm not a great soldering person, but I'm learning and I'd be game to buy a kit system and give it a go. I'd need people to help me in a crunch if I'm stuck. Hopefully, the forums and maybe a discord channel could help with that? Maybe there could be a youtube training video where somebody goes through the whole process of making one?

Edited by Danathar
ran it through grammarly ;)
  • Like 2
Link to comment
Share on other sites

Guest
This topic is now closed to further replies.
×
×
  • Create New...

Important Information

Please review our Terms of Use