Epic! 8-bitguy uses 1 weird trick to detroy rare prototypes!

[Mikoyan]

I think Terry would be more impressed with the Agon Light than the Commander X16, even though the former uses a Z80 instead of a 6502. It's way cheaper and much simpler than the X16. After all, a genius admires simplicity and an idiot admires complexity.

The problems with both are they're 8/16 bit systems (though there was a 32-bit Z80oid to build on). God said no paging, flat memory model, identity mapped and maybe said 64-bit.

Having said this, like most religions they're all paths back to the same central truth: "I like playing close to the metal like the c64 days". One does it with a simple graphics model, its own C variant, and lots of multi-core 64-bit power. The others do it by trying to be a neo-64.

And then there's the Mega 65, trying to recreate the original neo-64. Though as Bil Herd once said, "Nobody wanted that thing."

 

[AmpleApricots]

A chip being listed "in production" doesn't necessarily mean it's available to mere mortals, (often severely outdated chips are still "in production" in a way where the manufacturer makes very limited batches out of contract obligations, because he basically promised that this chip will still be available in 30 years or some such crazy number, most often to fulfill some government contract demand) especially at numbers a small setup like this one needs. You'd also be surprised how unwilling companies can be to sell their product to you in this sphere if you're below a certain threshold. I haven't looked into it at all so I might be dead wrong, but that in production yahama IC might be less available (esp. in needed numbers) than you think.

Making something like the x16 wasn't a good idea - if he wanted to make a commercial nostalgia project, he could have just built a hardware bus interface for VICE running on an arm SoC (like the c64 maxi but with real IO ports and a cartridge interface). Most of the motherboard/vera might as well be on a single chip as far as end users are concerned.

Yes. We're living in an age where $5 microcontrollers have more raw processing power than high-end 90s PCs. It gets harder and harder to justify to not just do everything in software when retro tech is concerned. I've seen a video on youtube of some ESP emulating a DOS PC, doing PS/2 input and VGA output in software via GPIO. As you said, for the user it basically does not matter and it solves a lot of problems. These systems are trivial to implement for anything just a little beefier in that sphere.

 

[PixarsTedKennedy]

I'm a little sad with how the X16 is ending up. I was too young to experience something like the C64 when it was new...but 500 dollar estimated cost, he said in his last video? I barely felt good about spending that on a PS5. I was hoping for a cheap, cool little computer that had a nice sized community that seemed fairly based, for the most part. Like that Pico emulator but without any trannies.

 

[AmpleApricots]

I'm a little sad with how the X16 is ending up. I was too young to experience something like the C64 when it was new...but 500 dollar estimated cost, he said in his last video? I barely felt good about spending that on a PS5. I was hoping for a cheap, cool little computer that had a nice sized community that seemed fairly based, for the most part. Like that Pico emulator but without any trannies.

I made my own Z180 based system (which sounds a lot more impressive than it is, but I'll never stop mentioning it because I consider the Z180 the superior choice for homebrew Z80-based projects) and there are also ready made kits out there if you're not sure where to start on such a project. They're cheap and the parts are still widely available. There's tons of ways to design hardware add-ons for such systems by yourself, easy to grasp even if you're not an engineer and you'll learn a lot, you can also make them look incredibly professional these days, having a professional PCB manufactured in china is super cheap. You might reach that $500 eventually, but not in one go and if you are just some dude who doesn't want to produce 10.000 units sourcing the proper vintage components is usually not a big problem. I even got an elusive speech synthesis chip for mine which name I'll not mention and see if somebody knows it.

This is just an expensive trinket for "retro consumers" who'll brag to each other how they own one and who will never use it for anything interesting or worthwhile, if they use it at all. (which really, is a huge if) Many such cases in this sphere.

 

[PixarsTedKennedy]

I made my own Z180 based system (which sounds a lot more impressive than it is, but I'll never stop mentioning it because I consider the Z180 the superior choice for homebrew Z80-based projects) and there are also ready made kits out there if you're not sure where to start on such a project. They're cheap and the parts are still widely available. There's tons of ways to design hardware add-ons for such systems by yourself, easy to grasp even if you're not an engineer and you'll learn a lot, you can also make them look incredibly professional these days, having a professional PCB manufactured in china is super cheap. You might reach that $500 eventually, but not in one go and if you are just some dude who doesn't want to produce 10.000 units sourcing the proper vintage components is usually not a big problem. I even got an elusive speech synthesis chip for mine which name I'll not mention and see if somebody knows it.

This is just an expensive trinket for "retro consumers" who'll brag to each other how they own one and who will never use it for anything interesting or worthwhile, if they use it at all. (which really, is a huge if) Many such cases in this sphere.

Well shit, I'll have to look into some of these kits. Any recommendations?

 

[Pee Cola]

Well shit, I'll have to look into some of these kits. Any recommendations?

Not a kit, but I stumbled upon a book from the early '80s called 'Build Your Own Z80 Computer'. It's very thorough, and even includes source code for a simple operating system written in Z80 assembler.

There's a simpler tutorial on how to build a Z80 computer from scratch here.

I wouldn't mind building a Z80 computer from scratch one day, but I'll probably just buy an Agon Light in the meantime to futz around with. They're cheap and there's a bit of a community forming around it already.

 

[AmpleApricots]

Well shit, I'll have to look into some of these kits. Any recommendations?

I'd look into things implementing the bus the RC2014 uses as there are a lot of available modules and homebrew stuff for it and you're not completely stumbling into from-scratch territory, which can be intimidating. If you google around a bit you can find quite a few kits implementing this bus in some form, all of them should be in the sub $100 range, some of them have versions of CP/M adapted to them. Be aware though that these are usually systems that are meant to be used with a serial terminal, so you'd mostly use them from your PC. There are expansion cards using vintage graphics chips though.

Otherwise, you might want to attempt to build a ZX Spectrum, which has all the fun parts like graphics and audio already and can largely be built from off-the-shelf logic. (The soviet bloc loved cloning it for this reason) I'm not entirely aware of whatever project is current there but some googling probably will lead you there too.

Otherwise you can go the from-scratch route as mentioned, the entry bar is a bit higher though. There's also the 6809 in that 8-bit sphere which is a more interesting CPU and less "dumb" than the Z80, but not as straightforward to program. There's also the 8088 but it's not as fun to work with as the other two tbh.

 

[Fcret]

David, you fucking retard, assembling this shit by yourself is easily going to triple the effective price of these units and either you or the users (or both) are going to have to eat that.

Is it because you hate China? There are experienced NA companies that'll do this for you at a fraction of what it's costing you now. Fuck man, you live in Texas - you could probably call around and get a lead on a maquiladora that'll do it cheap enough for you to still make a profit at the original price and they'll be glad for the business.

Not to mention that you and your manchild assembly line are probably going to do shit work and end up having to eat a shitload of returns when you inevitably produce tons of DOA units.

I'm a little sad with how the X16 is ending up. I was too young to experience something like the C64 when it was new...but 500 dollar estimated cost, he said in his last video? I barely felt good about spending that on a PS5. I was hoping for a cheap, cool little computer that had a nice sized community that seemed fairly based, for the most part. Like that Pico emulator but without any trannies.

The secret is that there's nothing particularly special about the C64 - it's an old bleepy shitbox which is only interesting if you explicitly have nostalgia for it or like the constraints of building things for old bleepy shitboxes. These systems are interesting in a historical context (which is why I enjoy retrocomputing youtube) but there's nothing all that special about using them nowadays.

Of course, the secret sauce these guys always talk about is the 'magic' of being able to program basic directly on the system itself. But none of them actually do this - they all use modern systems and emulators to write their code and then transfer them to real hardware. Because even back in 1983, no one wanted to write software directly on these microcomputers and most only did it because they were forced by financial considerations (and even small software companies would immediately upgrade to actual workstations the second they got enough cash to justify the purchase).

If you want to learn about how computers work though, building a Z80-based system is usually a good idea because shit for them is everywhere. Although if you don't really have any electronics or software experience, I'd recommend getting a modern SBC or microcontroller with some amount of peripheral support and just experimenting. You could build a pretty neat 'retro-style' computer just with an arduino and a few shields.

 

[Pee Cola]

If you want to learn about how computers work though, building a Z80-based system is usually a good idea because shit for them is everywhere. Although if you don't really have any electronics or software experience, I'd recommend getting a modern SBC or microcontroller with some amount of peripheral support and just experimenting. You could build a pretty neat 'retro-style' computer just with an arduino and a few shields

The Agon Light is ideal for this kinda stuff, as it uses a real Z80-based CPU (eZ80) and a barebones ESP32 for controlling video, IO etc, along with a set of GPIO pins. Firmware is programmable in C and the community is very active. From what I've read, at last count over 1000 Agon Lights have been made (mainly by Olimex and PCBWay), but this doesn't count the Agon Lights made by others from the plans on GitHub. Did I mention that was this only released at the end of last year?

Disclosure: I bought an Agon Light a few weeks back, but I'm yet to play around with it.

 

[PixarsTedKennedy]

The secret is that there's nothing particularly special about the C64 - it's an old bleepy shitbox which is only interesting if you explicitly have nostalgia for it or like the constraints of building things for old bleepy shitboxes. These systems are interesting in a historical context (which is why I enjoy retrocomputing youtube) but there's nothing all that special about using them nowadays.

Of course, the secret sauce these guys always talk about is the 'magic' of being able to program basic directly on the system itself. But none of them actually do this - they all use modern systems and emulators to write their code and then transfer them to real hardware. Because even back in 1983, no one wanted to write software directly on these microcomputers and most only did it because they were forced by financial considerations (and even small software companies would immediately upgrade to actual workstations the second they got enough cash to justify the purchase).

I can't argue that. I like to make Atari 2600/NES games, or try to, on my free time in 6502 Assembly but I recently watched a video of some guy doing it on an Apple II like it was 1979 and it looked like a huge hassle to me. I'd rather just code in VSCode, with syntax highlinging and debugging, while listening to youtube in the background. I think ultimately any retro computer would just collect dust on my shelf.

 

[AmpleApricots]

David, you fucking retard, assembling this shit by yourself

lmao, he does WHAT

I like the classic 8 bitters because of nostalgia and the feeling of being entirely in control and knowing how every single bit in that thing works. Also simple enough to just cook something up quickly. Most microcontrollers are not only a lot cheaper but also a lot more useful, faster and easier to work with. A lot of that nostalgia in my own design was to use ICs from back then and keep more modern things away. It's very tempting to use an ESP32 (or raspberry, or whatever) for supportive functions because it makes things easy, but it's very unfitting. It's like buying some 8-12 core PC monster with some 4090 GPU and having the OS run on the small ARM µC in your keyboard while it uses that PC as peripheral via USB to push 16 color graphics. It just doesn't make much sense and doesn't feel right.

16 bit is the realm where things start to be able to stand on their own feet, and early 32 bit processors could be comfy for daily usage if it wasn't for the internet. Also established systems from that time (e.g. System 7 era Non-Jobs MacOS, AmigaOS, DOS) have a big software library and engaging games that are still interesting/somewhat useful today. But 8 bit? Imagine having to use short variable names in BASIC to save memory. Like, one letter ones. That's what we're dealing with here. Just not very good. Memory access is shockingly slow so more complicated software (both writing and using) will always be torture. I think the biggest reason you don't see the retro affectionados flogging to processors like the 68k is because they're complicated enough that it becomes a day time job to do something on them, if you catch my drift.

 

[MembersSchoolPizza]

16 bit is the realm where things start to be able to stand on their own feet

I think it was Yatzee who said, at least of gaming, that the 16 bit era is when you didn't have to follow "These graphics look good" with ",considering". And it's true, to me.

 

[Fcret]

I think it was Yatzee who said, at least of gaming, that the 16 bit era is when you didn't have to follow "These graphics look good" with ",considering". And it's true, to me.

16-bit was also the era when you started to see multitasking, actual real filesystems, and displays that could actually render text as sharply as the high-end terminals that'd been used in the minicomputer scene. Like I could go back to an Amiga and probably not have too bad a time programming on it for an entire workday. I definitely could not program a full 8 hours on a C64.

 

[MembersSchoolPizza]

16-bit was also the era when you started to see multitasking, actual real filesystems, and displays that could actually render text as sharply as the high-end terminals that'd been used in the minicomputer scene. Like I could go back to an Amiga and probably not have too bad a time programming on it for an entire workday. I definitely could not program a full 8 hours on a C64.

Well, a few of the 8-bits could render sharp text... Even the C64 could, with one of two or three addon cartridges. Although that's all it could do, in that case... IRC it was limited to strictly text and strictly black and white, to do that.

The problem there was that most people didn't have monitors, or at least not good ones, and most people were connecting to televisions.

The rest of your points are fair, though.

 

[Susanna]

16-bit was also the era when you started to see multitasking, actual real filesystems, and displays that could actually render text as sharply as the high-end terminals that'd been used in the minicomputer scene. Like I could go back to an Amiga and probably not have too bad a time programming on it for an entire workday. I definitely could not program a full 8 hours on a C64.

Isn't the major factor there going to be the keyboard? The C64 used a TV instead of a monitor for cost reasons, there were actual monitors for it you could buy if you wanted, so the image quality was more a product of how home users employed it, rather than something inherent to the platform. Offices would have bought the proper monitors.

The 8-bit micros I've seen all had keyboards that make the garbage you'd get with a 2002 Dell look like IBM masterpieces (the ZX Spectrum didn't even have keycaps, you just squished directly into the rubber membrane), while the Amiga had proper linear switches.

Also, the Motorola 68000 in the Amiga was technically a 32-bit processor. For a 16-bit machine you'd be looking at something with an Intel 8086 or 8088 (IBM PC, Compaq PC), or the Z80000 Olivettis.

 

[ditto]

The only 8-bit micro I remember having a decent keyboard was the Apple II.

Lord British had a pretty boss development setup, at a time when even one floppy drive was a luxury. But notice the reams and reams of paper. Debugging was a bitch.

 

[Pee Cola]

The 8-bit micros I've seen all had keyboards that make the garbage you'd get with a 2002 Dell look like IBM masterpieces (the ZX Spectrum didn't even have keycaps, you just squished directly into the rubber membrane), while the Amiga had proper linear switches.

Sinclair keyboards were exceptionally bad. Atari ones weren't much better (the Atari ST's keyboard used a membrane with rubber plungers and feels like you're typing on bread dough).

The more expensive 8-bit machines (such as the Apple II and BBC Micro) actually had really good keyboards, especially the latter. I know they're virtually impossible to find outside the UK, but try and have a play with an Acorn Electron one day if/when you ever get to a computer museum that has one on display. It looks like a toy computer, but the keyboard is right up there with the IBM Model M - apparently Acorn spent way too much of the Electron's budget on the keyboard, and it shows.

16-bit was also the era when you started to see multitasking, actual real filesystems, and displays that could actually render text as sharply as the high-end terminals that'd been used in the minicomputer scene. Like I could go back to an Amiga and probably not have too bad a time programming on it for an entire workday. I definitely could not program a full 8 hours on a C64.

CP/M didn't have a real file system? OS-9 wasn't really a multitasking multi-user OS?

As for rendering text, I realise this is the exception rather than the rule, but Acorn's 8-bit machines were excellent in 80 column mode via monochrome composite, and did a decent job via RGB as well.

To say that all 8-bit machines were shit is a bit unfair, though I understand why someone would arrive at that conclusion if their 8-bit experience started and ended with the ZX Spectrum or C64.

 

[Fcret]

CP/M didn't have a real file system? OS-9 wasn't really a multitasking multi-user OS?

I'm talking about the mass-market consumer hardware that retrocomputing manchild youtubers are nostalgic for. Hence why I said "16-bit era" instead of specifically calling out hardware specs.

Although your CP/M example is a little egregious. It had the same flat file layout that every other shitbox of the era did. Being able to organize your files into hierarchies is what I consider to be the bare minimum for a file system to be tolerable to use (and no, user areas are not a substitute for real file hierarchies).

 

[Susanna]

I'm talking about the mass-market consumer hardware that retrocomputing manchild youtubers are nostalgic for. Hence why I said "16-bit era" instead of specifically calling out hardware specs.

Although your CP/M example is a little egregious. It had the same flat file layout that every other shitbox of the era did. Being able to organize your files into hierarchies is what I consider to be the bare minimum for a file system to be tolerable to use (and no, user areas are not a substitute for real file hierarchies).

To be fair outside the UNIX minis the standard at the time was saving to tape or floppy. Hard drives would take another decade to become commonplace in micros.
You don't really need filesystem hierarchies on a floppy, where the functionality the hierarchy gives you comes from just swapping floppies instead, and it would be pointless on a cassette, where you don't really have a file system at all, just timestamps for where each file begins.

 

[DagothUr]

16-bit was also the era when you started to see multitasking, actual real filesystems, and displays that could actually render text as sharply as the high-end terminals that'd been used in the minicomputer scene. Like I could go back to an Amiga and probably not have too bad a time programming on it for an entire workday. I definitely could not program a full 8 hours on a C64.

What really is the purpose of the 8-bit microcomputers? Obviously there's games and basic, or stuff for learning programming, but do they have any use besides that?