Mini Micro Fantasy Computer
miniscript.orgI’d love to see something like this but designed to run on esp32 or raspberry pi 2530. Either can handle basic HDMI and USB. Or a little <$100 laptop with a 7” display.
Easy to think raspberry pi, but with a full Linux you won’t get that intrinsic understanding that you fully control the hardware, you never control the “bare metal” unless you are a much more advanced user.
IMHO the feeling of not being in full control of your computing device is not a good starting point. I’m very fortunate to have started out on my 8kb BASIC machine.
There’s a wealth of retro machines out there that cater to this. A sample: https://www.olimex.com/Products/Retro-Computers/Neo6502/open...
https://www.olimex.com/Products/Retro-Computers/AgonLight2/o...
Or with keyboards: https://wildbitscomputing.com/
https://www.kickstarter.com/projects/spectrumnext/zx-spectru...
Or you can go mobile: https://www.clockworkpi.com/product-page/picocalc
https://andywarburton.co.uk/post/gr3ml1n-a-compact-handheld-...
And if you want a real challenge, the one euro computer:
https://www.olimex.com/Products/Retro-Computers/RVPC/open-so...
Something like the one dollar riscV but with an ESP32, hdmi, and an sd card slot. Pretty sure this exists somewhere. But it needs a good OS with a built in dev environment to be successful. I think I looked into if pico8 or something could work on the esp32, but it was too heavy iirc.
I have been thinking along those lines myself.
I have been playing around with a per scanline generated display on a rp2350 outputting to a tiny LCD. I think there's potential for some pretty fancy stuff on HDMI. A 2350 with PSRAM, HDMI connector plus a MicroSD for bulk filesytem, and USB for input could be quite a fun micro PC.
I would be tempted to make somthing that had a second RP2350 with its own PSRAM sitting unutilized just as a temptation to users to figure out how to get more out of the gadget and learn about different multiprocessing architectures.
One of these https://www.waveshare.com/core2350b.htm
With one of these https://www.waveshare.com/rp2350-matrix.htm
Mounted on top, and an HDMI connector squeezed in somewhere,
I am a bit reminded of what GeoWorks Ensemble managed on a 640k 8086. Theoretically you could make a tiny system like this do even more.
> I am a bit reminded of what GeoWorks Ensemble managed on a 640k 8086.
I was looking at similar recently for a project, and came across FrankOS: https://github.com/rh1tech/frank-os
Check out the Adafruit Fruit Jam, its got pretty much everything you need.
I really like the Fruit Jam. I was also excited about Frank's OS, I'm looking to see how hard it would be to port over.
Yes, they are more powerful than classical MS-DOS PCs, so there is plenty of juice in them.
There's PicoMite: https://geoffg.net/picomite.html
It's a BASIC interpreter/OS for the RP2040
You could just use RISC OS on the PI for BBC Basic.
There‘s LEDmeplay https://mithotronic.de/ledmeplay.php
Not sure on the performance but it might be possible to port this Mini Micro to those platforms.
Doubtful. Isn't Mini Micro build on Unity? That has much higher system requirements.
why not just use a vintage computer or game console then?
The main thing is video output. Even VGA is fading away now. HDMI is kind of what you need to be relevant to a lot of potenial users.
The machines were simple but the technology constraints of the day also made them very inconvenient. Lots of wires, bulky & slow storage, limited connectivity.
i haven’t seen hdmi on esp32. any example projects?
https://miniscript.org/files/MiniScript-QuickRef.pdf:
“A class or object is a map with a special __isa entry that points to the parent. This is set automatically when you use the new operator.
Shape = {"sides":0}
Square = new Shape
Square.sides = 4
x = new Square
x.sides // 4So
- Shape is a map (it is created using the syntax defined earlier, using a literal string as key)
- Square is a class?
- x is an object?
Or is this language prototype based? If so, why mention the word “class”? If not, isn’t it confusing to use “new someMap” to create a class and “new someClass” to create an object?
I also find it curious to see that division is defined on lists and strings. What would that mean?
Edit: reading https://miniscript.org/files/Strout_iSTEM-Ed2021.pdf, it is prototype based. That’s interesting for a teaching language.
Yeah, I guess it’s prototype-based and the authors meant classes are indistinguishable from objects. And they all are just special cases of map.
It's so odd that the only nontrivial example code in the paper is completely buggy. The find longest common prefix function of a list of strings fails (try ["a", "bc", "ade"]).
I was a bit confused until I realized that https://miniscript.org/ isn't the same programming language as https://bitcoin.sipa.be/miniscript/.
This looks interesting, and it’s good to have alternatives to Lua in the embedded space, although MiniScript is an awful lot like Lua on first glance (e.g. using keywords instead of brackets to end loops and conditionals).
The things which slightly rub me the wrong way:
• C++ instead of C. There’s still places where one has a pure C language project.
• cmake instead of Posix-compatible make. This makes the project depend on a non-standardized tool with only one implementation.
I’m curious what advantages MiniScript has over Lua, though—when I had the problem of “let’s use a tiny embedded language” [1], Lua5.1 made the most sense to me:
• No need to worry about a Javascript, Moonscript, Kotlin, or Go language port—people have already made all of those. [2]
• People already made libraries I needed (e.g. a “spawner” library so I could run Stockfish from my Lua script)
• It’s a standard embedded scripting language which people are more likely to already know (e.g. people making Roblox games already know Lua)
• It compiles in a pure C environment.
[1] https://samboy.github.io/MaraDNS/coLunacyDNS/
[2] It would be nice to have a Rust port of Lua
Apparently it’s high-level only, i.e. no underlying machine instruction set or addressable memory.
So... It's an interpreter (together with a virtual filesystem and some utilities) packaged into a program with a graphical display window? Still good for lots of interesting uses, I suppose, but surprising. Since it's introduced as a "virtual computer", I thought underneath the hood it would be emulating a machine. Then people, if they wanted, could tinker a level deeper than the scripting language, write an assembler etc.
That's how most of these fantasy machines are. Most people are only going to want to use a high-level language so it makes more sense this way.
True , but if it had a real ISA underneath then we could write custom emulators for them.
Also check out this one. It is a real physical computer: https://www.francksauer.com/index.php/micro-8
Free but not Open Source? Did I miss that?
> Free but not Open Source? Did I miss that?
The miniscript language itself is MIT License:
https://github.com/JoeStrout/miniscript
The Minimicro code doesn't seem to have any license in the repository or code:
So Open Source but not Free (Libre).
For miniscript: the Free Software Foundation considers the MIT license (which they call the ‘Expat License’ to distinguish it from the ‘X11 License’) to be ‘free’ (and GPL compatible), but not ‘copyleft’.
https://www.gnu.org/licenses/license-list.html#Expat
For minimicro-sysdisk: I am suspicious that the author just forgot to include a license. Their other repos are mostly MIT or ‘The Unlicensed (also ‘free’ but not ‘copyleft’), and some have licenses added after creation. Suspicion is not something to be legally relied on of course…
That's usually called "source available", since most people, including some governmental orgs around the world, already follow OSI's definition for "open source".
Am I the only one that hates that programmers now have to also be lawyers?
What do you mean now? The peak of debating the merits of varied FOSS licenses must have been 20 years ago.
Mini Micro seems to be built on Unity. The MiniScript portion of it is open source https://github.com/JoeStrout/miniscript but the version packaged for use by Unity costs some money. I can't tell if the people behind MiniScript are the same people behind the Mini Micro.
I don't get why this kind of projects need :
- a manual
- an installer
when you have Web pages can now
- be offline (PWA)
- be responsive and run on pretty much any device
- run pretty much anything thanks to WASM but anyway already have JS/HTML/CSS as bare minimum
- can have the instructions AND the runtime on the same page, on any device, instantly
- can connect with physical hardware, see recent https://hacks.mozilla.org/2026/05/web-serial-support-in-fire... or even with APIs.
> Mini Micro is a neo-retro virtual computer. Learn to code, play some games, and join a friendly, enthusiastic community of hobbyists!
Apparently not that friendly on HN :-D
What's unfriendly in questioning the delivery mechanism or design choices?
I can think a project is cool or useful without thinking it's perfect. In fact ALL projects I encounter, including mine, are like that, imperfect. I think native is a barrier to entry so IMHO it should only be relied on when it's truly necessary. I'm just stating my opinion and others are welcomed to disagree and explain why they think so.
It's a way of expressing things. "I don't get why you do X" generally comes across as "it's not smart to do X".
Similarly, "this kind of projects" sounds derogatory, as you don't specify what kind you're talking about, which implies "the kind of projects where the author didn't think things through thoroughly".
Also "AND", why do you need to capitalize here? Cause you want to show that your idea is a much better one.
I think you know pretty well that you were criticizing the project in a way that was not very friendly, as in: if an actual friend of yours had made the project, you wouldn't have used the same wording.
Web browsers suck as an app platform. Give me native any day, personally.
Looks cool. I most enjoyed the zombies game someone uploaded on itch.io. One thing to note is that game speeds feel very fast to me. I barely did anything in the asteroids game and the others also seem to run quite fast. It could be just me.
I wonder how hard it would be translate this to Dutch. I would like my kids to start experimenting but that’s a bit impractical if they need to learn English first..
Ik denk dat Claude dat zo voor je doet in een paar minuten tijd.
I wonder why all these easy-to-learn languages use indentation to denote scope, not something like curly braces. Isn't it actually harder to explain?
Fifth grade teacher here. Significant whitespace is a major reason I prefer Python for teaching programming.
1. I want kids to indent their code anyway; they may not realize it (or won't admit it), but this makes the code much easier for them to read. Kids will not do this unless they have to.
2. Unbalanced brackets are a major source of mistakes and confusion for my students. Relying purely on indentation resolves this problem—at the real cost of introducing indentation mistakes, but since I want kids to indent their code anyway, this is okay.
By the way, an adjacent recommendation is to configure the editor to indent with tabs instead of spaces (regardless of how you feel about tabs vs spaces in production code). Otherwise, kids will invariably end up with lines indented by 3 or 7 or some other wacky number of spaces. If possible, highlight the tabs in a different color so the kids don't use spaces by accident.
Interesting point. I wonder if "easier for them to read" is too simple. I took "read" as in "read words" or "read a book". But "reading" a program is not I think the same as reading words. Reading words could be this:
for i = 0 i < 10 i++ if i = 7 printf("hello 7") else printf("who are you");
But with a more pictorial presentation, it is easier to read the program.
for i = 0 i < 10 i++ if i = 7 printf("hello 7")
I'm just wondering - if we had a more pictograph based programming language would it be easier to understand?
Sounds like Scratch:
This is the job of a tool like go fmt. Obviously, it’s good discipline to indent, but I wouldn’t choose this as the deciding factor for picking a first programming language.
Formatters and linters fix the mistakes made by people who know what they're doing. They do nothing to teach someone how to do something for the first time in a way that supports comprehension, only regurgitation.
go fmt can fix #1, but not #2, and won't work if #2 is causing problems.
The alternative is what the Visual Basic editor does: It indents automatically; not only when pressing return, but whenever the caret is moved to another line. This means you virtually can’t fail to notice that you missed an END IF or similar, because subsequent lines will be indented as part of the IF body.
It looks like MiniScript uses the keyword "end" followed by another keyword to denote the end of a specific type of block.
From the Quick Reference guide here:
https://miniscript.org/files/MiniScript-QuickRef.pdf
"Indentation doesn't matter (except for readability)."
while everyone already pointed out that this time it's not the case, I want to literally answer the question you asked
"easy-to-learn languages" use indentation because otherwise newbies would not indent at all
I you try teaching programming, you'll find that indentation is one of things students "optimize out" - it is not important to the program, it is opposite of lazy and it's not noticeably harmful on the tiny scale of programs you learn programming from
Indentation discipline only starts to matter when you need to work on the same code for quite some time and code itself takes a lot of space - the "read more then written" situation. And most study paths do not encounter this regime
Easy-to-learn and built-for-learning-good-practices are two different things;
I think your point is perfectly correct but it's mostly about the second one
I think indentation is more intuitive. Even people using languages that use braces or similar usually use indentation to make code readable. If doing that you end up explaining both ideas (use braces and indent).
I get why people like indentation for this. I don’t think there’s a right or wrong answer and it’s a matter of personal preference.
That said, my preference is curly braces (or whatever) because I’ve found indentation is often a bother. Yes, most of the time you use indentation together with braces, but not every time. There are many occasions where code is clearer without (or with custom) indentation. Furthermore, indentation-based parsing makes experimentation and finding issues more difficult. Sometimes you need to extract a small part of a larger block to bung in a REPL or something and now you’re fighting with stupid errors because of formatting, adding to the frustration.
Regarding intuitiveness, for beginners I have some doubts it makes much of a difference, and if it does I also doubt indentation wins. If you know how to write (which is a prerequisite), you know what parenthesis and quotation marks are, you understand they encapsulate something separate from the rest. Indentation is a different concept.
I get why people blame indentation like this. I don't think it's right or wrong to ignore the tooling that directly addresses minor issues with indentation or matching braces honestly.
That said, my preference is to use the tools built into my editor and available on the CLI or web to assist and fix formatting and syntax. You get instant feedback on incorrect formatting, and I generally find that synthetic scope mistakes (regardless of method) are eliminated.
I think indentation tries to step towards an outline, but without the item indicators we’d use in an outline format. This might actually be a selling point for something like YAMLScript.
There's an important form/function distinction here, though. Indentation is useful for human readability, but braces function to give unambiguous direction to the compiler or interpreter. I think conflating these two different purposes together is a mistake: you shouldn't risk altering or breaking the logic flow of a program simply by adjusting its visual formatting.
The fact that we use whitespace for layout is precisely why it's a bad idea to assign it semantic value. I'm a fan of both braces and semicolons for that reason.
I think this is probably correct for an experienced programmer but incorrect for someone who is new.
I'm not sure I understand what you mean. How would the friction inherent in conflating layout and semantics together depend on the experience level the programmer? Different programmers might have different ways of dealing with that friction, but I'd think its existence would be a property of the language itself.
The form/function distinction you're making requires the ability to hold two parallel representations of the same code in your head—the visual representation (what it looks like) and the syntactic representation (what it means to the parser), and to know that they're related but different. This is a higher level skill.
When you're starting out, the best form to express to other humans is probably the one you're expressing to the computer. This isn't literally true—I don't think beginners should write in assembly—but it's true enough that they probably shouldn't mess with indentation beyond what would naively follow from bracket placement.
> The form/function distinction you're making requires the ability to hold two parallel representations of the same code in your head—the visual representation (what it looks like) and the syntactic representation (what it means to the parser), and to know that they're related but different. This is a higher level skill.
My supposition here is that the threshold of skill needed to understand the form/function distinction here is significantly lower than you expect it to be. In written natural language, people don't typically attach semantic meaning to things like indententation, kerning, letter spacing, text alignment, font size, etc., and usually distinguish style from semantics intuitively without having to apply any conscious decision rules. Distinguishing form from content is something we do by nature.
I'd expect that at least those already familiar with natural writing in most modern languages will have a bias toward expecting that whitespace does not have semantic value, while punctuation marks do. Conversely, I'd expect recognizing cases where indentation does control logic flow to be what requires a higher skill level.
The language (MicroScript) doesn’t require indentation, it’s only used for readability, like in BASIC, FORTRAN, PASCAL, and similar languages. Blocks are delimited by key words (“end if” etc.).
It makes sure the thing you use to judge scope (indentation) matches the think the computer uses.
That's a fair point for students, but as a beginner who simply wants to tinker with fun stuff, you can go very far without knowing of a program stack.
I think I had the wrong audience in mind
Why not for 3 eur buy some basic arduino or other tiny hardware to tinker with and for another few eur, tiny i2c/oled display, wires and set of basic switches? You start programming with option to expand to the larger project in the future. You have constraints of real device, community is much larger and there are more learning resources.
Because those don’t boot into a fixed interactive programming environment with a BASIC-like language and REPL to easily do simple things on the same screen and using the same keyboard you also use for programming. Your proposed setup has more complexity and is less intuitive for a learner.
For starters, there is way more friction both in buying hardware and waiting for it to arrive and developing on real hardware in general.
I agree however that it's super cool to have real hardware to run this on.
Because moving a sprite is much more exhilarating than blinking an LED.
Well how about moving a sprite by blinking a few leds?
Because those require you to get to grips with “a tiny i2c/oled display, wires and set of basic switches” when you’re interested in coding, not hardware.
I think that's really cool. I wonder when this started development?
Shame there isn't a Raspberry Pi version available.
It is built in Unity which doesn't let you build for ARM Linux without paying an unknown amount of money.
Couldn't someone just rebuild the source code for Raspberry Pi?
What source code?
There‘s Pico-8 for the Pi!
Is this a similar project to the existing Pico8?
Probably more like Picotron. https://www.lexaloffle.com/picotron.php
This feels nostalgic!
Ah the nostalgia
really like the creativity, cool stuff
Only virtual? That is sad!