Change of product direction, good and bad news!

[AndyMt]

7 hours ago, Carl Gundel said:

Some of 8BG's games also do this, so why can't developers be encouraged to write games that work on both machines?  After all we do have access to an X16 emulator, so even if the X16 isn't available when the X8 is, there's no reason why software can't theoretically be developed for both, and published for both?

For sure this is possible, but - it takes away time from my budget to actually develop something. I would certainly only develop for one of the 2 platforms and not go into the hassle to do ports. I'd probably also wait until there is a clear indication of which platform "wins". Maybe I'll then not develop for any of the 2 at all because my focus moved somewhere else.

7 hours ago, StephenHorn said:

But there will be one X16, and maybe the X8. And the install base will be in the hundreds, maybe (maybe!) the thousands. This is much less incentive for wide support, especially if folks write software to be released for free, as opposed to software they intend to sell. Or, to the extend that wide support is encouraged, it will be by developing to the minimum spec shared between the two machines.

I fully agree. Let's assume the X8 gets the same VERA interface as the X16 and is tuned down to 8MHz. I'd probably just develop to the minimum specs. For the 2 games I've released so far I pushed the X16 as far as I could. I've used almost all the VRAM (256 color bitmap graphics, lots of 256 color sprites) and also plenty of banked RAM (for music and sound). My platform jump & run game in the works goes even further: sound track running in parallel, 256 color palette switching to show up to 512 colors at once, partial palette shifting to achieve "water flow" effects etc. I'm not sure what I'll do with it now - probably just wait and see what happens.

1 hour ago, Stefan said:

I do not believe in hiding the differences between X16 and X8 behind API layers. 8 bit computers running at 8 or 12 MHz need all computational power they have if we are to make great programs. Even if there were such an API many assembly programmers would avoid it to gain performance.



If the X16 and X8 are incompatible, the programs written for them will be so too. If both the X8 and the X16 are released, developer time will be divided between the two platforms. 

Here I also agree. If the VERA interface is different, then the software will be incompatible. Or the effort to make it compatible will take away performance, which could be too much.

[Amy Worrall]

1 hour ago, Stefan said:

I do not believe in hiding the differences between X16 and X8 behind API layers. 8 bit computers running at 8 or 12 MHz need all computational power they have if we are to make great programs. Even if there were such an API many assembly programmers would avoid it to gain performance.

One of the reasons I love the BBC Micro is that it _does_ hide many things behind API layers. It was one of the only computers of the era with a fully featured operating system with lots of API calls. That's what allowed things like running software on an external second processor with minimal modification.

Not necessarily commenting on whether the same tradeoff would make sense here, but it's an interesting system to look at if you want to see how such an OS could work in the 8 bit era.

[zapposh]

[VincentF]

20 minutes ago, zapposh said:

It's not possible with the X8's VERA implementation, we got a 256 bytes window that is just impossible (or too expensive) to do on real hardware for the X16 ?

[Robinkle]

My first choice:

The Phase-2 X16 with FM chip in a mini-itx format. Preferably a complete system assembly with the keyboard included.

My Second choice:

The Phase-2 X16 with FM chip in a mini-itx format, bundled with some stickers and IO shield, so that I could install it in any SFF-case of choice. Keyboard included.

About the X8. I would buy it too. And add free open source case design for 3D-printing, that users can download. It’s a perfect product for those who want a base model.

As for Phase 1, it seems it might be expensive and demand a ton of work. It would only make sense to make it DIY kit, because this product really appeal to people who would like to tinker with it anyways. A bundle of parts, case and keyboard seems like a great idea.

 

[BruceMcF]

1 hour ago, VincentF said:

It's not possible with the X8's VERA implementation, we got a 256 bytes window that is just impossible (or too expensive) to do on real hardware for the X16 ?

Masking out a single page at a fixed address for the I/O page is bad enough ... having a "roving" page to access the Vera not be practicable. There's only so many layers of chip select logic they can have and still work with the through-pin, 55ns SRAM that they have available.

[Stefan]

If doing code for both X16 and X8, I would for performance reasons first look into creating a set of macros rather than using a common API.

As to bank switching, such a macro could on the X16 just select the RAM bank.

On the X8 that macro could load a virtual memory file (for instance numbered 0-255) into memory area $a000-$bfff. However, first the current content of $a000-$bfff need to be saved to its virtual memory file.

Maybe similar solutions are available for VERA access. I haven't given that much thought yet.

[kelli217]

When I first read this post and its comments, it was still all only on one page. As of right now when I'm starting to type this response, there are 10 pages. Whew.

But since we're all offering our opinions, even people who have only been lurking up till now, I think that there should be a Phase 1 with discrete through-hole components and the same number of expansion slots as there already are on the Proto 3 board and it should come exclusively as a kit with a keyboard but without a case. It's MicroATX, there are a bajillion cases out there. You could even go buy the original version of the case that @Perifractic modified. Just, you know, independently.

I think that there should be a Phase 2 that also offers expansion slots but fewer of them, perhaps 3, with a cost-reduced motherboard, maybe in a mini-ITX form factor. Integrate as much as you can, reduce chip count, etc., and use the lessons learned from that and from the X8 as the jumping-off point for the Phase 3. Sell it fully assembled with both keyboard and case.

The Phase 3 will be the 'laptop' version. No one expects to be able to stick a full-sized expansion card in a laptop. Any expansions will have to come through the VIA port. Did I mention that I am hoping all of the phases will still have a VIA port? Yeah. Wimodems and printers and other odds and ends make retro systems more fun for people who are new to the retro scene to use in the modern era. It should also be sold fully assembled and with a keyboard and case, or a keyboard that is also a case, a la the RP400.

Every X16 should have 2MB RAM as standard. Three versions of the product are enough; no point in complicating things by having different RAM specs on top of that.

Regarding the X8: Less capable (less RAM, less VRAM, no FM chip [please keep the FM chip on the X16]), less compatible (changes in VERA addressing, ports), less retro (USB ports instead of PS/2 and SNES)... but ready RIGHT NOW. Sell it, ASAP, at a profit, and reinvest those profits into X16 development.

Also, the X8 is probably going to be subject to the actual Osborne Effect rather than the reverse. As the X16, and in particular the Phase 3, comes nearer to fruition, the X8 will likely go the way of the dodo. Given that, and its potentially low price even with the profit factored in, meaning that it won't deplete people's bank accounts, and its existence shouldn't cannibalize the X16's sales too much. I've bought multiple Raspberry Pi versions, at least one of every generation, simply because they're so inexpensive, even though I rarely do anything with them.

And the VERA addressing might not even be that much of a hurdle to compatibility. If the 256-byte window of shared RAM happened to default to starting at $9F00, wouldn't that make it possible for the same first few bytes to be accessed the same way as shared RAM as they would as I/O space on the X16?

Last but not least: Those of you who have said that they would rather donate directly rather than buy an X8, great! You don't have to wait for the product to be available for sale nor for David to set up a specific donation infrastructure; you can donate right now, to his Patreon! I did! Here's the URL: https://patreon.com/8bitguy1

[Felipe Balbi]

Personally, I'd like to see X8 and X16 on the market as kits :-) If you launch the final product with a case or not, it's less relevant :-)

[Jeroen]

After a cooldown of a few days, I have changed my mind.

My initial response was no/no/DYI, driven from emotion primarily. 

I have built my own 6502 board from scratch. In fact, two versions of it.

While developing software, guess how many times it was actually interesting that it runs on discrete ICs. Not many. After implementing an XModem bootloader, I didn't even have to touch the EEPROM at all. Finally.

I did however quickly felt limited by the I/O options that I built in. It has a PS/2 keyboard and serial I/O, but video is still lacking. Then I got engaged with the X16 and built software for the emulator. My 6502 boards haven't been powered on for some time now.

In retrospect, even without knowing all the details from the X8, when this would have been the first option presented to me a year ago, branded as X16, I would have been sold.

Sure, bank switching, I/O slots and a dedicated sound-chip are interesting future options. But given the power of the VERA, I'm not really interested.

So please, deliver the X8 as the first option (with the cool keyboard ? ). I'll rewrite my sokoban clone when I have access to the emulator. Indeed doesn't sound like much work.