Like many people my age (ahem!) the Commodore 64 was my first computer and I loved it and it's responsible for my long software development career.
But I don't want that Commodore 64 today.
I want the Commodore 64 of 2025. A machine where middle schoolers can learn the basics of programming while having fun with graphics and sound. Maybe even have a simple 2D gaming engine built-in. I don't know. I want the spirit of the Commodore 64, not the actual machine itself.
> I want the Commodore 64 of 2025. A machine where middle schoolers can learn the basics of programming while having fun with graphics and sound. Maybe even have a simple 2D gaming engine built-in. I don't know. I want the spirit of the Commodore 64, not the actual machine itself.
Exactly. This is what I think every time I see one of these old revival projects. I don't want a Spectrum, C64, Atari ST, etc...we have those, they're fairly easy to acquire and renovate. And are more than capable of being run on a FPGA. And there are dozens of projects built around the same old 6502, Z80, etc.
Stop locking your perspective into the 80s to try to recapture that nostalgia.
No, give me a new "hobbyist" computer in the spirit of those days. Throw an ARM m-series/RISC-V/etc on it with some custom blitter/vdpu and sound ASICs and 512MB of RAM. Give it some easily accessible programming environment on ROM, with an option to baremetal with ASM, C, etc. Add a few slots that are MMAPed in. And let the hobbyist field run wild.
> No, give me a new "hobbyist" computer in the spirit of those days. Throw an ARM m-series/RISC-V/etc on it with some custom blitter/vdpu and sound ASICs and 512MB of RAM. Give it some easily accessible programming environment on ROM, with an option to baremetal with ASM, C, etc. Add a few slots that are MMAPed in. And let the hobbyist field run wild.
Isn't a pi awfully close to that, at least in spirit? For my 10 year-old's science project, I bought a trio of interesting sensors off amazon, showed her a diagram of the GPIO pins and a diagram of the sensor pins, explained how to map between the two, and had her draw it out with colored pencils.
Then I burned a fresh raspbian image onto a sd card, connected a keyboard, mouse, and tv, and helped her figure out how to read the GPIO pins in python. The vibe of the whole thing felt a lot like the old things we had in the 80s and early 90s, but more accessible because I didn't have to deal with weird serial/parallel junk or with putting together a PCB for the slots.
It does sound like this crew harbors ambitions of moving past nostalgia to embracing that spirit a little more, but I don't personally feel like that's lacking in the Pi ecosystem, at least...
> The vibe of the whole thing felt a lot like the old things we had in the 80s and early 90s
Except back then you were doing all the discovering and figuring stuff out without any help from your parents. At least that's the way it was in my circle of friends. When a 10-year-old is able to do everything you did without any assistance from an adult, then we'll have the spirit of the Commodore 64.
I appreciate that. I would add that she might be able to do everything I did (with my 8-bit micro when I was 10) with the pi I set her up with. The thing I was impressed with about the pi was how little I had to help her interface with sensors that touched the real world, like light sensors and temperature sensors. The GPIO there is much more accessible to her than the user port on the C64 or the expansion slots on the Apples were to me.
I had to help find the references and explain some of them to her, but I've been pleased with how little help it's required from me. Probably only a little more help than I needed to learn how to connect cables and format a floppy before I ran off on my own as a kid. But she's getting more independence at the end of my help than I got with the hardware of my day.
It’s similar on the hobbyist interest side, sure. But generally gathers more from the hardware hobbyists than software. This is because the Broadcom SoC is somewhat awful to do bare metal on (mostly due to anything beyond a framebuffer being mostly locked off), so you’re limited to using Linux for full capabilities.
Nothing wrong with that, but definitely a different environment than something like BASIC as your bootrom, or direct hardware poking you’d get from the old hobbyist machines.
I have always wanted to learn to build a (relatively modern) computer like the one you described:
- Some 32-bit CPU, whatever, anything that is a bit easy to program through C/ASM, just to make sure there is no weird kirks.
- Support keyboards, displays, mouse, etc., just the usual ones. So a lot of drivers.
- Some 256MB - 512MB memory should be good enough.
- Has an OS, some programming languages, some tools, a good editor, etc.
This is like the Ben Eater 8-bit computer in adrenaline. It is probably a LOT of work just to figure out how to source the correct components, and build the thing, then a LOT more work to write drivers for them, and MORE work to write OS and compilers and tools for them. We can't use Linux because it has memory protection all over the place. We need something that newbies can poke and peek into, and simply reset the CPU if something is wrong, just like the micro computers in the 70s/80s. We DO want capable compilers and interpreters (e.g. C/Python) and good tools (like, some editors that have good auto-complete at least).
It's a bit like building a pad or a mobile phone, but without all those commercial consideration. Nowadays, to build a pad or a mobile phone, if I'm not mistaken, one simply push Android into the chipsets and call it a day, which is not what I want. But what I want probably doesn't make $$ so no one is going to do it.
> Support keyboards, displays, mouse, etc., just the usual ones. So a lot of drivers.
You wouldn’t even need that if you released it as an all-in-one like the hobbyist machines of yore (C64, speccy, etc). Just give it a relatively decent cherry MX built-in with swappable keyblocks. Add a USB hub for HID devices and include a basic HID driver, and include HDMI.
The hardware is the easy part. But if you want a selection of software such as exists for the C64, that takes a lot more time and of course there’s no guarantee that it will ever come to pass.
If you’re trying to make yourself into a massive company off of it, sure the interest definitely doesn’t exist and you should forget it.
If you look at the amount of interest projects like the openpandora, the retro rereleases, etc get; there’s plenty enough for something to hack around on. You would just have to release it as a side/passion project with modest revenue projections.
That's what I want too. But I fear that even if such a machine existed, the spirit wouldn't be there. Just like no one forms bands or makes personal websites anymore.
I've been into 'retro computing' continuously since the 80s and there have been a variety of interesting retro focused machines like this launched over the last 15 years. They are mostly hobby projects but some have scaled up through crowd funding to be marginally successful. A big success in this market would be shipping a thousand units but few reach that level. I think the reason is that despite quite a bit of interest the market is fragmented into different groups with differing primary goals:
* Capture "the spirit" of an all-in-one, simple computer that boots to an accessible language.
* Recreate an actual 80s computer via software, FPGA emulation or compatible CPU. A major sub-group is those wanting support for physical 80s media which can include disks, cartridges and even tape.
Within these major groups are a variety of different requirements. A big one is whether the machine must support modern displays (HDMI) or authentic retro displays (CRTs). If you don't have a CRT then HDMI is a requirement, however inserting what was originally low-res analog composite video into a hi-res digital container involves some significant trade-offs and design complexity. It's not trivial or cheap to do well with high-quality and high-compatibility. Then there are those who split on whether modern connectivity and conveniences like Wifi, Ethernet and SD card media are mandatory, nice-to-haves or definite should-not-haves. Of course, those conveniences aren't much use without sufficient CPU power and resolution to support a modern browser and OS capable of reading modern media which involve more cost and potential compatibility issues.
The great thing is that those who are retro-interested now have a lot of good options ranging from OG hardware to software emulation, FPGA systems and all-new designs. My advice is to be clear on what experience you really want, the specific traits you care about and the various trade-offs and challenges those entail before diving in.
I think the challenge is computers these days can do so much that tinkering with something like this no longer feels futuristic or cutting edge like a C=64 did back then.
Sure, a modern SBC repackaged in an all-in-one case can be a good option for most of the first group I mentioned. However, out of the box it doesn't boot to BASIC like an early 80s home computer so you'd need to configure a custom boot and choose a language. And even within that group there are some for who the simplicity of limited graphics, resolution and other capabilities is an important element, so they'd probably want to go further to something like auto-booting directly into a software emulation of an 80s computer.
For those who care about running retro software titles, software or FPGA emulation would be the minimum, however that doesn't help those who want support for retro physical media or peripherals, CRT displays or highly accurate emulation. You may not care about those things but for a significant part of the addressable market for machines like this, one or more of those things are a high priority. Which was my point, the retro-interested market looks pretty large but when you zoom in, you realize it's fragmented and that makes it challenging to design a device that appeals to enough buyers. Since the RPi 400 seems perfect to you, you probably don't care about it looking like a C64, however the folks behind this new device paid a lot to buy the Commodore brand name and logo (supposedly up to a million dollars). And that's before the plastic moulds, retro keyboard etc. For you, that may all be wasted money and effort. For others it might be a big part of the appeal.
It's certainly possible to make a retro computer that has HDMI, Wifi, Ethernet and SD cards and gets on the modern web but also has analog composite, S-video, and component output for CRTs, along with having a cartridge port, floppy drive, serial or parallel port and which runs period software with cycle-accuracy. However, that device is probably going to cost close to $1,000 and take a fair bit of time and expertise to create. I'm a hardcore retro-enthusiast and I'd certainly consider spending that kind of money but most people probably wouldn't. Worse... I'm not even in the market for such a wonder-machine because I already own over a hundred different models of retro computer along with analog composite, S-video and composite CRTs to use them (almost every 80s model of Commodore, Atari, Apple, Amiga, Sinclair, Amstrad and Radio Shack along with other more obscure brands).
The Mega65 was a stab at this idea - a self contained modernized version of the old 8-bit computer while trying to maintain backwards compatibility with C64 programs.
Which is based on the 6502-compatible 65C816 but used a simple banking scheme instead of the broken 24-bit address space that chip natively supports (no 24-bit index registers) The way video memory works in it is really clever and lets it really surpass 1980s machines in many ways.
which is priced right though it doesn't have the keyboard and instead based on the eZ80 which really does extend the Z-80 with 24-bit registers so that you can use all the RAM easily.
The guys behind Olimex are top chaps!
Their products are not only affordable, also OpenSource as well!
Nothing good came so far by the C= or Amiga copyrights...
Are you remembering the same BASIC that I’m remembering? With fixed line numbers? Refactoring was near impossible. Spaghetti code was par for the course. You were forced to build everything twice or thrice if you wanted a final product that wasn’t embarrassing.
Just not having line numbers cooked in is a total game changer.
I had a C16, and later a C128D (and therefore a C64). The C64 had the most limited BASIC of them all, with v2.0. So many advanced features were hidden behind PEEK/POKE. With BASIC 3.5 on the C16, they added the RENUMBER command, which made refactoring much easier, and it had a SOUND command, despite having a much simpler DSP implementation. The C128 had BASIC v7.0, which included a sprite editor, and much more advanced SOUND command.
If they upgraded the C64's BASIC to 7.0, that would already make a lot of things much more accessible.
> With fixed line numbers? Refactoring was near impossible. Spaghetti code was par for the course.
Yup, even for the old 8-bit computers FORTH would've been a lot more elegant than BASIC. But back in the day BASIC came with highly valued conveniences such as a soft-float implementation that meant support for the "desk calculator" use case was available out of the box, and users could just go on from there.
There are many flavours of BASIC; some with line numbers and GOTO/GOSUB, some without line numbers and real subroutines. FreeBASIC for example doesn't have line numbers. It's similar to Fortran or COBOL, both of whom also started out with line numbers and GOTO-oriented programming. But no one uses those languages like that any more.
No idea what this BASIC looks like though; the documentation seems chaotic and hard to access. Just saying that BASIC does not necessarily mean the BASIC that you or I used back in the day. I also agree that you can do a hell of a lot better than Python to introduce programming to kids.
Plenty of elementary and middle-school kids learned LOGO which is a vastly more elegant language than BASIC. (It's essentially a LISP with different syntax, the one thing that's inadvisable about it for modern purposes is its use of dynamic scope.)
I think understand of programming as a sequence of commands is necessary before you go to abstractions. BASIC's GOTO is a feature, not a bug: that's how assembly work.
As an adult I'd strongly prefer Lisp over Basic. But I appreciate Lisp or lambda calculus elegance after I spent some time with more basic kind of stuff. Going to higher abstractions right away won't have the same effect
Python would be the wrong choice. BASIC is the ultimate "batteries included" language. No weirdness hidden behind dynamic types and you can make full games with graphics and sounds (in most implementations) without installing a single external library. BASIC is feature-poor by design and beginners can quickly fully master it. And there are classic variants without line numbers (no need for GOTOs) for those who hate them.
IMHO, humanity is yet to build a better beginner language, at least if you look at the late BASICs such as QuickBASIC.
Besides being integral to the experience / faithful to the prototype C65, it is remarkably reliable storage media that's well-suited for this class of machine, which have very few KBs of RAM.
In that regard, I almost feel like a new Atari 800-series would be better.
The C64 had good graphics and excellent sound but so much of it was behind a brick-wall learning curve of poking. Atari's native BASIC at least provided some rudimentary access. You want something where the user can get a win on day 1, or it's getting buried in the closet with the rock tumbler.
Or maybe if they packed in a super-extended BASIC ROM. But pretty quickly you end up wanting something with more modern flow control and structures, maybe closer to "Qbasic with sprite commands" and then you're probably demanding more than what can be reasonably asked of a 6510-class CPU.
>The C64 had good graphics and excellent sound but so much of it was behind a brick-wall learning curve of poking. Atari's native BASIC at least provided some rudimentary access. You want something where the user can get a win on day 1, or it's getting buried in the closet with the rock tumbler.
This was not my experience at all.
I started with an Atari 400 when I was 11, and there was little to no documentation about the machine. I took a BASIC programming class taught on a Vic 20 when I was 10 years old. But then I saved up and got an Atari for my 11th birthday. I cut my teeth on Atari Basic, but it seemed so limited to me. I wanted full access to all of the machine but there was simply no documentation my 11 year old self had available, the internet was not a thing, and there were no books about it at my local library in the early 1980s. The BASIC cartridge didn't really offer much in the way of using the machine to its full capacity.
Then I got a C64 and everything changed. The C64 manual I had included BASIC programming, as well as how to program the sound chip to make simple music and sound effects. There was a complete memory map of the C64 including all the chips, the VIC, the SID, the CIA chips - everything. All the pinouts for all the I/O ports were mapped out.
And in the back of the manual were all the opcodes for the CPU, with detailed information about how they work, A, X, and Y registers, status registers, program counter, address modes, etc, etc... it was absolutely fantastic how much was documented, literally everything about the computer. The book even had a full schematic of the C64 in the very back as a large fold-out poster. Atari had nothing like this.
It was like night and day compared to the very closed Atari platform. I quickly ate up BASIC on the C64 and moved right into assembly language, when I was 14. I got very into the European "demo scene", while being a teenager in America in the mid/late 1980s. Atari 8-bit just seemed like a toy compared to what was going on with the C64 in the world at the time.
I agree. I recently started exposing my kids to programming and I chose a C64 emulator. The BASIC REPL is so simple/limited that it doesn’t overwhelm the kids with irrelevant syntax (those can come later, if they are interested). The fact that a 1..100 loop can introduce a noticeable delay. You can literally see the computer working. Primitive graphics and sound provide immediate feedback which makes learning engaging and fun.
Tulip CC is the modern C64. The software is mostly fleshed out for audio apps right now, but feature for feature, it's doing the same kinds of thing, with the same eye to budget:
I would argue that we have that today in the form of the Apple MacMini (https://www.apple.com/shop/buy-mac/mac-mini) -- the Commodore 64 was priced at around $300 in 1984 or $950 in 2025 money. The basic model MacMini comes in at $600 today. For that you get a fully Unix system with a full development suite to build desktop/mobile/Unix apps out of the box. Every development platform is available (C/C++, Python, Swift, JavaScript, Java, Rust, etc.) on it. The expansion system is extremely simple also (USB C/Thunderbolt). Not to mention it is much smaller than a Commodore 64 and arguably more user friendly.
This was great at the time, I feel it should of come with a 'computing partner doc' like a piece of paper with how to backup and restore with a second sd card, and links to stuff about censorware and what people think about certain places attached to the internet..
I saw one of these at a pawnshop with a $250 price tag and I think a chrome book can be had for less.
not sure what the current availability and pricing of PI type things are these days - but the kano kit was a perfect fit at the right time back then.
Some of it is astonishing given the limitations of the machine.
The technical mastery and creative problem solving going on here is what keeps the demoscene alive. The C64 is still going strong, new productions come out all the time.
And I can't think of a better machine for a kid today to start out with than an 8-bit machine. If they learn assembly language on a C64, they will have a far better understanding of how computers actually work than anyone taking a class in web or crypto or whatever they teach kids about computers today.
I think part of that spirit was not having the internet. You had the machine, a couple books and a handful of magazines. But more importantly an unstoppable desire to figure out how it worked. The internet would have been a distraction.
(disclaimer, I have maybe a bit too much of this retro hardware.. unhealthy amount, ahem).
Raspberry Pi 400 / 5 would be that.. but, in reality and pragmatically since it needs to be something widespread I'd argue it's actually a browser. In Chrome you hit Ctrl-Shift I and console is right there.
Is this all just nostalgia?
Nostalgia is one of our two core pillars - alongside modern innovation. Like yin and yang, these forces balance and strengthen each other in that retro • futurism approach.
The commercial power of nostalgia is real - and it will help fuel and fund the development of modern, forward-facing products in turn. It’s a symbiotic cycle: retro inspires, modern sustains. Commodore isn’t returning. It’s evolving, with purpose.
They also mention that they explicitly wanted this - a state-of-the-art reimplementation of the old C64 - to be their first released product, which makes some sense. It's also the product where their Commodore trademark - the real value behind this new effort - is most relevant, shifting away from the old pattern where random products would be "Commodore" branded, with no real link to the company's history or to any plausible "retro futurism" vision.
Agreed it looks like slop, and it's IMO a bad sign. I think a big part of the appeal of old computers is the fact that they're simple enough for a single human to completely understand.
Generative AI is a black box that's impossible to completely understand. Using generative AI signals to me that whoever did it probably doesn't find any inherent value in understanding things, and sees understanding only as a means to an end. Old computers have little practical use, so this leaves nostalgia as the main appeal, and nostalgia has less stringent requirements.
To kids these days, a laptop with a something like Roblox Studio on it is pretty magical. Most of their "laptop" exposure is with Chromebooks that only run school stuff, everything else is on consumption-focused tablets and phones that don't leave much room for creativity, or even basic customization.
I have a Commander X16 and my kid had fun doing that Basic infinite maze program, but after that, my dream was over :)
But I don't want that Commodore 64 today.
I want the Commodore 64 of 2025. A machine where middle schoolers can learn the basics of programming while having fun with graphics and sound. Maybe even have a simple 2D gaming engine built-in. I don't know. I want the spirit of the Commodore 64, not the actual machine itself.