I'm not trying to be a turd... please take this as constructive criticism, in case you haven't thought about these questions...

I agree that it'd be really neat to have a little board with an integrated CPU, VGA and ideal DOS sound, but it will still have the limitations of a similar system made from old parts. The main differences being that the board will be proprietary and not as expandable (can't swap CPU, VGA, sound, etc.), it will likely cost far more than a pile of old parts (the components needed for a DOS gaming PC are not rare or expensive... it's not like you're integrating 3dfx hardware), and the elephant in the room is that it doesn't really have the draw of "vintage" hardware since it's a modern custom device. And this isn't even touching on the idea of turning into an entire console (which implies ease of use, customer service... and profit?).

Also, if a nearly perfect OPL3 clone is sufficient, why not one of the nearly perfect OPL3 emulators, like the one used in DOSbox (after some tweaking?) or Nuked OPL3? I keep reading that people find them indistinguishable from a real OPL3.

For the SoC choice, how would you handle speed sensitivity? What is going to manage the speed of the CPU to ensure that every game runs smoothly, but doesn't have glitches? DOSBox does a good job of this automatically, and it can be adjusted manually with little effort. With physical hardware, you either have to spend time with each game, experimenting and tweaking (disabling caches, messing with FSB or multiplier, etc.) or you have to build different systems for different types of games. With a pre-built DOS console, obviously the second option won't work, so is the user going to be responsible for doing all the tweaking for each game? Would someone have to program something that would launch games with tweaks set for each one? Who is going to program that and then make configurations for all the games, assuming it is possible with the hardware chosen?

And in the end... how is this experience more authentic than running an emulator, aside from the physical chips doing the processing? DOS-era PCs were never known for having console like ease of use, so having a launcher that is carefully set up to optimally run specific DOS games is not going to set off the nostalgia center of any geeky brains that I know. Currently, retro PC gaming enthusiasts likely fit into one of two camps:

1. Someone wants to use an old computer to play old games (either because of nostalgia or an interest in learning about them), and they get to experience what it is like to use the original hardware to run the software they want to run. They can have as much authenticity as they want. Real OPL3, ball mouse, noisy hard drive and a CRT... or CQM, USB mouse on an adapter, SSD and an LCD with some converters? On top of this they get options that nothing outside of real actually vintage hardware can give them (the look of a DOS game on a CRT, their favorite game music playing on one of the many fascinating wavetable and FM implementations on various sound cards, scanlines from 3dfx SLI, the look of 24bit-16bit dithering in Glide on a Voodoo 3, the sound of Aureal A3D in real time, the amazement of seeing an old game looking better than ever running on the very latest hardware that can run it... or the ability to play games that only work in Windows 9x for that matter).

2. Someone simply wants to play old computer games and they don't so much care about all the hassle of building and operating an old computer (can you blame them?). They can either buy from GoG, or they can download DOSBox, read the existing documentation online about how to make that work, and then they can play whatever they want. They can get nearly perfect visuals, amazingly accurate sound (even MT32 or Roland GM\GS synth are possible), the ability to run it on nearly any computer (and any display + peripherals), and there's a massive community to support either option.

If there is going to be a third option... who is looking to game on a device that brings most of the quirky downsides and complication of gaming on an old computer but has limited upgrade options, software that few people have experience with (since it likely won't run MS-DOS and will have other custom software), and none of the authenticity of running an old computer... but they still are picky about hardware FM synthesis (but not necessarily a real OPL3!) and that the CPU is 486-like but is a totally modern design.

Again, what is the goal? Who are we designing this for? 😮

Ozzuneoj wrote on 2020-12-12, 03:15:

Again, what is the goal? Who are we designing this for? 😮

People who think it's cool for some reason. There's really nothing more to say, I guess.

Jorpho wrote on 2020-12-12, 03:25:

Ozzuneoj wrote on 2020-12-12, 03:15:

Again, what is the goal? Who are we designing this for? 😮

People who think it's cool for some reason. There's really nothing more to say, I guess.

If people think it's cool, they can certain pursue whatever they want. People make modern clone devices all the time, to fill a really tiny niche.

If a lot of resources are going to be tied up to make a commercial product however, the only semi feasible product I can imagine that would be close to this would be a prebuilt x86 mini-computer that is set up for GoG Galaxy (but only the old games?). At least they have patched many of the issues with popular older games. Titles that generally only work in Windows 9x or XP have work-arounds included to make them work. DOSBox is included (and just works) for DOS games, regardless of the game's requirements. You can use modern controllers (with some messing around), modern displays, etc..

This solution ensures compatibility, requires little effort from the customer, offers a good (sometimes better, in the case of games that have been updated) experience with few compromises, and if it was put into a miniature beige tower and sold with a beige mechanical keyboard and a retro themed GoG Galaxy launcher, it would have tons of retro appeal to the masses, regardless of what was inside it.

Problem is, everyone can get almost the exact same experience with one of the hundreds of millions of modern computers laying around unused. They can likely do this on the computer they already have.

So, even that product would be a stretch, and probably never break even before being canned.

If someone just wants to build a fully custom semi-retro-hardware-based-gaming-PC for the fun of it, I'd love to see it, and I'd certainly appreciate the work that went into it, but it could never realistically be a product.

AFAICT GoG do far less work than people give them credit for. Usually their "fixes" for windows games are just bundling freely available directdraw replacements and other community-sourced solutions. If a game has issues with DEP (like Diablo and Warcraft 2 did when they were first released) their solution is to make the end-user add a security exception for the game - despite that not being an available option for current 64-bit windows platforms. Their support for MacOS is even worse, not bothering to update many titles for 64-bit compatibility leaving people unable to play games they had purchased - just as if they'd purchased the games when they were originally released, so I guess it's an authentic experience!

jmarsh wrote on 2020-12-12, 04:21:

AFAICT GoG do far less work than people give them credit for. Usually their "fixes" for windows games are just bundling freely available directdraw replacements and other community-sourced solutions. If a game has issues with DEP (like Diablo and Warcraft 2 did when they were first released) their solution is to make the end-user add a security exception for the game - despite that not being an available option for current 64-bit windows platforms. Their support for MacOS is even worse, not bothering to update many titles for 64-bit compatibility leaving people unable to play games they had purchased - just as if they'd purchased the games when they were originally released, so I guess it's an authentic experience!

Yeah, don't get me wrong, there's a reason I'm here and I have 1600+ posts almost entirely in the Marvin forums... I prefer tinkering with real hardware most of the time.

Still, GoG is undeniably more accessible than what we do here.

I'm veeery slowly working on a VLB video card, which has DVI output (which would work with many devices with HDMI inputs). My hands are full with other stuff though and it will take years before it is done...

Ozzuneoj wrote on 2020-12-12, 03:15:

I'm not trying to be a turd... please take this as constructive criticism, in case you haven't thought about these questions...

I appreciate the constructive criticism. Thank you for taking the time to write down your thoughts.

Ozzuneoj wrote on 2020-12-12, 03:15:

I agree that it'd be really neat to have a little board with an integrated CPU, VGA and ideal DOS sound, but it will still have the limitations of a similar system made from old parts. The main differences being that the board will be proprietary and not as expandable (can't swap CPU, VGA, sound, etc.), it will likely cost far more than a pile of old parts (the components needed for a DOS gaming PC are not rare or expensive... it's not like you're integrating 3dfx hardware), and the elephant in the room is that it doesn't really have the draw of "vintage" hardware since it's a modern custom device. And this isn't even touching on the idea of turning into an entire console (which implies ease of use, customer service... and profit?).

Can't argue with anything you said here. I will point out, though, that there's a market for the NuXT, even though that's pretty far from vintage anything, save the CPU itself, and is more expensive than cobbling together an XT clone oneself.

Ozzuneoj wrote on 2020-12-12, 03:15:

Also, if a nearly perfect OPL3 clone is sufficient, why not one of the nearly perfect OPL3 emulators, like the one used in DOSbox (after some tweaking?) or Nuked OPL3? I keep reading that people find them indistinguishable from a real OPL3.

I agree that Nuked OPL3 is an excellent OPL3 emulator, far better than many hardware clones of the mid 90's like CQM or ESFM. I only mention the CMI8330 because there are some who suspect it actually does integrate an authentic OPL3, just like the YMF719 or YMF744, but was never advertised as using a licensed core from Yamaha. I have considered and rejected the idea of running Nuked OPL3 on a cheap ARM Cortex-M or similar microcontroller - the cost wouldn't be much different compared to a YMF262 or YMF289, and using software emulation deprives this project of a reason to exist, even if that reason is somewhat irrational.

Ozzuneoj wrote on 2020-12-12, 03:15:

For the SoC choice, how would you handle speed sensitivity?

You are probably aware of the range of slowdown options out there, from tweaking multiplier and cache settings to TSRs. As a console, each game needs a profile to correctly map its inputs to the controller, correctly manage conventional memory, and set CPU parameters.

Ozzuneoj wrote on 2020-12-12, 03:15:

Is the user going to be responsible for doing all the tweaking for each game? Would someone have to program something that would launch games with tweaks set for each one? Who is going to program that and then make configurations for all the games, assuming it is possible with the hardware chosen?

Thus far, I imagine a rolling release with a growing library, where the first batch of game profiles are the most popular titles with the least challenging tweaks, and then moving to more complicated setups from there. I would welcome community contributions, and there would be a public bug tracker.

Ozzuneoj wrote on 2020-12-12, 03:15:

And in the end... how is this experience more authentic than running an emulator, aside from the physical chips doing the processing? DOS-era PCs were never known for having console like ease of use, so having a launcher that is carefully set up to optimally run specific DOS games is not going to set off the nostalgia center of any geeky brains that I know.

This is exactly why I started the thread. I find this idea fascinating, so am I the only one? It occurred to me that there is a way to have console-like ease of use for DOS games and still use real hardware and cover almost the entire library with a single hardware configuration. So, if that is in fact possible, the question I'm trying to answer is whether the idea is worth pursuing. I get the sense that you think it isn't, and that doesn't offend me. I really want to know how people react to this idea.

Tiido wrote on 2020-12-12, 13:02:

I'm veeery slowly working on a VLB video card, which has DVI output (which would work with many devices with HDMI inputs). My hands are full with other stuff though and it will take years before it is done...

If anyone can do it, I'm sure you can. It's tangential to this thread, but if I were tasked with such a project, I would start with a laptop VLB video chip with a parallel pixel bus output intended for an LCD like the WD90C24A, wire that up to a DVI/HDMI transmitter (maybe directly?) like the ADV7511W, and go from there. It's probably not actually that simple, but maybe it is.

Don't know if that's what you were already thinking.

Tiido wrote on 2020-12-12, 13:02:

I'm veeery slowly working on a VLB video card, which has DVI output (which would work with many devices with HDMI inputs). My hands are full with other stuff though and it will take years before it is done...

Fine.
Which graphics chip do you consider?

On another note kinda related to all this, yesterday I ordered an Intel Atom D525 development board. Seller still has some left. I made an offer of $75 from the asking price of $99 and the seller accepted.

Has PCI and ISA and from what I could gather there should be support for ISA DMA.

I like collecting development boards. They often have better features than the stuff that actually gets made and released in the public channel.

I still want to get my hands on a Vortex86 DX development board with PCI and ISA.

LightStruk wrote on 2020-12-13, 05:00:

If anyone can do it, I'm sure you can. It's tangential to this thread, but if I were tasked with such a project, I would start with a laptop VLB video chip with a parallel pixel bus output intended for an LCD like the WD90C24A, wire that up to a DVI/HDMI transmitter (maybe directly?) like the ADV7511W, and go from there. It's probably not actually that simple, but maybe it is.

Don't know if that's what you were already thinking.

My approach is FPGA based, so no existing chips get used. I'll implement all the important DOS standards and maybe XGA to gain hardware acceleration in windows (until I learn to write windows drivers). And there will also be a built in scaler so that monitor/TV can be fed with its native res always, and you get to choose exactly how the resolution needed to be shown gets scaled up and if framerate conversation is used or not.

Anyway the approach you described should work, and as simply as that, as long as the pixel data is always in the same format which it should be. There will be difficulties legally getting the HDMI chip though, unless you shell out for the license... that's one reason why I dislike HDMI, and would love to support DP except all DP chips are also HDMI capable and subject to HDMI licensing 😒

Since you took the time to reply, so will I. 😀

Again, I'm thinking about this as a finished, marketable product with enough happy customers for you to call it a success, because this seemed to be where you were going with this in the OP. As a thought experiment or a limited run DIY curiosity, with no intended financial gain, it is more feasible.

LightStruk wrote on 2020-12-13, 04:31:

Can't argue with anything you said here. I will point out, though, that there's a market for the NuXT, even though that's pretty far from vintage anything, save the CPU itself, and is more expensive than cobbling together an XT clone oneself.

That's a fair point. It does exist, and it is\was a product. The NuXT is an incredibly small niche product though. I would love to know how many they have sold at $340 (they're out of stock too). I think once the computer is assembled, the NuXT does provide a fairly authentic experience for those that would normally buy an XT clone and then purchase a bunch of modernized parts to make it work (so they don't need rare and expensive EGA monitors, XT keyboards, MFM Drives, RAM expansions... etc.). I think NuXT's main purpose is to replace all of the rare (and possibly broken) old parts that you need to build an XT clone. Some people may just want the experience of building a simple computer, using their modern peripherals (and any old DOS compatible VGA monitor...) and then playing games from the 80s. Pretty much anything pre-1990 will work on the NuXT with very little configuration required, using whatever version of MS-DOS the owner wants.

I find the NuXT very cool, and I'd be interested in playing with one if they weren't so expensive, but personally I wouldn't have much of a real use for it. I prefer to use the old keyboards, the old monitors, the old drives, etc. Again... I don't know how many they've sold, but I doubt it's in the thousands. Also, an XT board is far simpler than anything 16bit or 32bit. The work that went into that had to be considerable, but a 486 would require significantly more. There have been other threads here where people proposed building modern 486 boards, and I don't believe any have been made. And that's just a board... not a console with custom software and implied capabilities that users would expect to work (console-like game compatibility, modern game controllers, wifi, HDMI with DOS support, etc.).

LightStruk wrote on 2020-12-13, 04:31:

I agree that Nuked OPL3 is an excellent OPL3 emulator, far better than many hardware clones of the mid 90's like CQM or ESFM. I only mention the CMI8330 because there are some who suspect it actually does integrate an authentic OPL3, just like the YMF719 or YMF744, but was never advertised as using a licensed core from Yamaha. I have considered and rejected the idea of running Nuked OPL3 on a cheap ARM Cortex-M or similar microcontroller - the cost wouldn't be much different compared to a YMF262 or YMF289, and using software emulation deprives this project of a reason to exist, even if that reason is somewhat irrational.

I mentioned that only to point out that even enthusiasts at times don't notice the difference between software emulation and a real OPL3. If you're going to use hardware, use a real chip. But the sound likely won't be discernibly better than, for example, a modern PC running DOSbox. If a real OPL3 is a major selling point, the existence of emulation makes it harder to sell the product to anyone other than purists who want real hardware... which they likely already have access to.

LightStruk wrote on 2020-12-13, 04:31:

You are probably aware of the range of slowdown options out there, from tweaking multiplier and cache settings to TSRs. As a console, each game needs a profile to correctly map its inputs to the controller, correctly manage conventional memory, and set CPU parameters.

Yes, this is basically what I was thinking. Someone would have to do that work, which isn't trivial.

LightStruk wrote on 2020-12-13, 04:31:

Thus far, I imagine a rolling release with a growing library, where the first batch of game profiles are the most popular titles with the least challenging tweaks, and then moving to more complicated setups from there. I would welcome community contributions, and there would be a public bug tracker.

This is feasible but makes it hard to sell a product unless it's super compelling, priced well and has a guarantee of lots of support. Keep in mind, the people who fund the project will be in the red while trying to build this customer base and community. If your customer base is also narrowed down to people who want exactly this product, but are also willing to wait for the catalog of games to be fleshed out... it's getting smaller.

LightStruk wrote on 2020-12-13, 04:31:

This is exactly why I started the thread. I find this idea fascinating, so am I the only one? It occurred to me that there is a way to have console-like ease of use for DOS games and still use real hardware and cover almost the entire library with a single hardware configuration. So, if that is in fact possible, the question I'm trying to answer is whether the idea is worth pursuing. I get the sense that you think it isn't, and that doesn't offend me. I really want to know how people react to this idea.

The idea of such a thing is certainly fascinating, and I would be interested in it's progress... I'd follow it daily! But the interest of thousands of enthusiasts doesn't pay off your debts, which would likely be considerable after paying someone to design a 486 board, obtaining all the components, designing a case for it all, programming the software to make old games work better than in DOSBox (because there's no point if this can't be achieved)... and this only touches on the basics. Modern peripherals, wifi, HDMI, an operating system, customer service, marketing... yikes.

You could do a crowdfunding thing and end up going insane trying to live up to people's standards on the budget you've been provided...

But in the end, as mentioned, who is it for and how many of those people are there?

Also, I just discovered the PC Classic:

https://www.pcgamer.com/the-pc-classic-is-a-t … -for-dos-games/
https://unitetechno.com/dt_catalog/pc-classic/
https://www.youtube.com/watch?v=6f5PIrUT_Ts
https://www.reddit.com/r/pcgaming/comments/9w … tem_that_comes/
https://www.facebook.com/PCClassicUnite/?ref=page_internal

Functioning prototype shown, likely uses DOSBox or something similar for emulation (I didn't watch the whole video), got some interest from the press... and hasn't had an update since February (or any new videos since 2019). The company appears to have experience with other products already.

Not sure what happened to this, but it's worth mentioning.

Ozzuneoj wrote on 2020-12-13, 21:28:

Again, I'm thinking about this as a finished, marketable product with enough happy customers for you to call it a success, because this seemed to be where you were going with this in the OP. As a thought experiment or a limited run DIY curiosity, with no intended financial gain, it is more feasible.

You're on the right track. I'm not delusional; in my least plausible dreams this project could turn a small profit. It's far more achievable if I can figure out how to get it made as a passion project instead of a business.

Ozzuneoj wrote on 2020-12-13, 21:28:

There have been other threads here where people proposed building modern 486 boards, and I don't believe any have been made. And that's just a board... not a console with custom software and implied capabilities that users would expect to work (console-like game compatibility, modern game controllers, wifi, HDMI with DOS support, etc.).

Indeed, and I have participated in at least one of those 486 board threads. In this case, my intent is to use a highly integrated SoC, either vintage like a Geode, or modern like an Atom or AMD G-Series APU, to greatly reduce the number of components and the complexity of the board. Because this is a console with literally one stable configuration, I don't have to deal with design details like variable power demands, expansion of any kind, most of the external ports, etc.

Ozzuneoj wrote on 2020-12-13, 21:28:

The idea of such a thing is certainly fascinating, and I would be interested in it's progress... I'd follow it daily! But the interest of thousands of enthusiasts doesn't pay off your debts, which would likely be considerable after paying someone to design a 486 board, obtaining all the components, designing a case for it all, programming the software to make old games work better than in DOSBox (because there's no point if this can't be achieved)... and this only touches on the basics. Modern peripherals, wifi, HDMI, an operating system, customer service, marketing... yikes.

Don't forget that I have to succeed at getting VGA mode 13h (320x200@8bpp 70 Hz) through an HDMI cable and displaying on a recent TV! Some combination of upscaling (to 480p if nothing else works), Variable Refresh Rate, dealing with the HDMI licensing mess, figuring out whether I need an FPGA or if I can source an off-the-shelf HDMI transmitter that fits... there's a lot of unsolved engineering challenges even before I get to a hypothetical business part.

Ozzuneoj wrote on 2020-12-13, 21:47:

Also, I just discovered the PC Classic... Functioning prototype shown, likely uses DOSBox or something similar for emulation (I didn't watch the whole video), got some interest from the press... and hasn't had an update since February (or any new videos since 2019). The company appears to have experience with other products already.

Yes, I'm aware of the PC Classic. It's assuredly an ARM SoC running an emulator, just like the NES Classic, SNES Classic, Playstation Classic, et. al. I actually think the PC Classic's biggest challenge is input - what kind of device can you ship with the console that will play a bunch of games that required either as little as the arrow keys and spacebar or as much as a joystick, throttle, rudder, and dozens of keyboard keys? If their demo from 2 years ago is to be believed, their device won't even ship with its own gamepad, and will expect the user to supply a keyboard and something like an Xbox One gamepad.

I do have a gamepad design, and it's the most achievable thing that I have planned for this project. It's not just another dual-stick + d-pad + 4 face buttons and triggers design that every console since the PS2 has shipped with. If I can get the gamepad built and working, which requires some non-trivial engineering (PCB, microcontroller programming, bluetooth, chassis design, battery management, and more), then I can test my first assertion. I'm claiming it's possible to have a remappable input device that works on the couch and can play elaborate DOS games like Wing Commander, X-Wing, Mechwarrior 2, and Descent as well as it plays mouse games like Warcraft 2 and early FPSes like System Shock.

Even if the console is a pipe dream, I can ship the gamepad on its own.

I've thought about this for a long time.......but since it's fun....here's my thoughts.....

- Let's start with the CPU, I know of one nobody here has mentioned - Vortex86 - which is a 486-like x86 compatible CPU - www.vortex86.com - except it supports DDR3, up to 256MB of RAM, and is already in use for Embedded apps. It's basically a SOC 486 with some capabilities above a regular 486 DX based system. - www.vortex86.com

- For modern data storage, we would likely want to utilize some form of solid state, inexpensive, and easily swappable storage such as an SDCARD

- for graphics, we would need to offer multipke options, as not everybody would be putting this thing through a 4K HDMI only Television, we'd need at least composite and vga to offer the ability to connect it to a vintage display, or teat it as an actual desktop DOS machine if wanted.

- There would need to be some form of attaching older storage to the device - probably via USB - to install your data onto it off original discs/diskettes.

- Sound - for the best experience - would mean having PC Speaker, Tandy 1000 3-voice, and at least Adlib with 16-bit SoundBlaster digital audio, and have all tha tplay stable.

- Networking - now here's where it gets really difficult. If we want networking, most people have WiFi, this means we need DOS to be able to connect via WPA or WPA2-PSK - which as one who has 3 NEC Versa Laptops with DOS WiFi that I have to tether through a cell phone hotspot to use such technology - is impossible given what's available currently. It would mean whatever version of DOS we were using would mean we would need to find a way to support at least WPA-PSK over WiFi in DOS, which means creating a more robust WiFi subsystem than an Orinoco or Aironet chip with a mere packet driver and mTCP.

- Once all this is done, then we have O/S. FreeDOS is definatley a choice, and would be my first choice, but being DOS it would require the user to actually have some skill in actually USING DOS. An ideal solution would be a menu system that uses no or little RAM on boot off a ROM and then have that load the games off the HDD and be able to dynamically configure the DOS box to behave with said piece of software. Which at that point, you might as well be messing with RetroPie on a Rpi4 2GB like I already do because it's going to be every bit as insane and fiddly. Sure, there could be a preset of popular profiles for popular games - but not everyone will be running just Doom, Duke Nukem 3D, Skyroads, Lemmings, Sim City, and Monkey Island - which are all VERY easy to setup and get working - you might have something like Ultima 7 that uses it's own DPMI or the 7th Guest that expects a CD-ROM on startup - meaning a custom SHSUCD configuration to get it working off an ISO file the user may or may not have to rip themselves. What about booters too? Like Microsoft Adventure, or Donald Duck's speedboard race.

- Then there's people who'd want to run Windows on the machine. I've used HX in FreeDOS - or well....tried - and until that particular DPMI/Win32 loader program gets better support for things like Diablo, Streets of Sim City, or Postal, and easier for a non-technical user to configure, it's going to be a problem.

But I see what they say about it not being feasable. A PC is not like a games console - you can't just slap a few games on it and expect them all to work properly. I mean, just alone, as a retro-PC user myself, THIS is what I have to deal with...

- Special boot parameters for Ultima 7 and Serpent's ISle because they have their own DPMI and need at least 560K of the 640K base DOS RAM free
- The Secret of Monkey Island 1st VGA edition I have needs a patch applied for the SoundBlaster driver or you get a garbled mess when using Adlib
- Some games need set blaster=A220 I5 D1 P330 (or whatever your settings are) to work, some don't, so there's some RAM freed there
- I've got a pile of games that I use SHSUCD to load the ISO files for rather than dragging around nearly 30 year old CD-ROMs to play (Killing Moon, 7th Guest...etc)
- I "NEED" Windows to run things like Postal, DIablo, Beachhead 2000, Microsoft Golf 3.0, and so on, HX is just not up to snuff yet from what I can tell
- I need a special configuration if I want to be able to connect to the wired network, or to WiFi, and on WiFi, I can't use WEP (router does not support), WPA (DOS does not support, nor the packet driver), let alone WPA-2 PSK.....so I'm stuck either on a wired network, or tethering through my cell phone's hotspot
- Certain games run terribly/too fast on hardware above a 6Mhz 8088 - that's why I have a Tandy, a 286, 3 different 486s (2 laptops), 2 Pentiums (1 laptop) - because some stuff runs better than others, and it's not all based on clock speed of CPU or the graphics card either.

Now imagine some nostalgic layperson trying to do all that when the only things they know how to do with a computer is Facebook, Youtrube, Twitter, Outlook, Word, Skype, and "run Windows Updates". They might as well go fork out $45 for a reasonable old x86 box and start tinkering.

i'm not holding my breath for this, theres a better chance of somone coding a front end for dosbox that would work seamlessly on a console. which is on the the other hand could be profitable/feasible and workable. code an appstore for such front end that can do in app purchases from GOG or somthing then you have somthing u can actually work and make money from by taking a certain of money percentage from app sales.

Is there really all that much in DOS gaming that appeals to a broad range of people? I really can't believe people would readily embrace paying any sort of premium for four-color CGA with PC-speaker beeps.

I expect some sort of replica Amiga would be a much bigger draw, especially considering that there wasn't nearly as much variance in Amiga hardware as there was with PCs, and I can hardly imagine there are many Amiga games out there with superior DOS ports. This article suggests that the lack of a source for new custom chips is a major obstacle:
https://hackaday.com/2020/06/16/why-you-proba … a-anytime-soon/

So here's an idea I had...

Taking advantage of the ISA compatible x86 SOCs is really interesting, and I think using them to build a compact retro-computer would appeal to a broader audience than just the people who want to plug it into their living room TV. If the game console idea isn't more practical than a Raspberry Pi, how about making it something fun to build and tinker with by making it modular?
My idea is to create a standard similar to PC104, but optimize it to reduce costs as much as possible by using non-standard parts (like, say using cheap 2.54mm headers and sockets instead of ISA edge connectors), and focusing on modern I/O. The costs will never compete with Raspberry Pi, but my idea is to build the initial console from modules that adhere to this new standard, opening the door for future modules to meet a wider variety of needs. Maybe create a standard SBC based on a Vortex86 SOC and Vortex86VGA with HDMI output, USB input, removable flash storage, and then a couple of daughter boards (or a backplane made from cheap components with expansion boards) with accessories like sound and ethernet, all with modern I/O options. Keeping it standard could open the door for other people to create compatible hardware. For those who want to connect existing ISA cards, an adapter could be created.

I don't really know if this idea is any more feasible/practical/interesting than the original DOS gaming console idea.

johnnycontrario wrote on 2020-12-15, 06:44:

So here's an idea I had... […]
Show full quote

So here's an idea I had...

Taking advantage of the ISA compatible x86 SOCs is really interesting, and I think using them to build a compact retro-computer would appeal to a broader audience than just the people who want to plug it into their living room TV. If the game console idea isn't more practical than a Raspberry Pi, how about making it something fun to build and tinker with by making it modular?
My idea is to create a standard similar to PC104, but optimize it to reduce costs as much as possible by using non-standard parts (like, say using cheap 2.54mm headers and sockets instead of ISA edge connectors), and focusing on modern I/O. The costs will never compete with Raspberry Pi, but my idea is to build the initial console from modules that adhere to this new standard, opening the door for future modules to meet a wider variety of needs. Maybe create a standard SBC based on a Vortex86 SOC and Vortex86VGA with HDMI output, USB input, removable flash storage, and then a couple of daughter boards (or a backplane made from cheap components with expansion boards) with accessories like sound and ethernet, all with modern I/O options. Keeping it standard could open the door for other people to create compatible hardware. For those who want to connect existing ISA cards, an adapter could be created.

I don't really know if this idea is any more feasible/practical/interesting than the original DOS gaming console idea.

That probably could be and it could compete with the Arduino/Rasberry Pi hobbyist market. And it would probably fall nicely between the two - being as Arduino is a very basic microcontroller for simple things, and Pi is a64-bit ARM based Linux computer - a Vortex86 sort of thing would fall in the middle - an x86 FreeDOS Hobbyist computer. We could just attach a simple GPIO setup like the Pi has and then people could make their own ISA/PCI/VLB/etc adapters and "hats" for it and put it in whatever form factor they want - whether that's in a tiny box velcroed to the back of the TV (like my RetroPie is), or a full blown retro gaming laptop like I know they make parts to make with the Pi units.