I want to build something similar but with bare metal Lua as an educational computer for kids. For all of the STEM toys on the market none have the magic of a 8 bit machine that boots to Basic. Except in this day and age Basic should be replaced with Lua.
I love the idea of this, but some time definitely needs to be invested in reducing costs. $340 for a plastic housing and some screws? That's ludicrous.
I bet you could CNC some injection molds the cost of 2-3 of these chassis. If someone could do that and sell the housing for ~$30 each, this would be an impressive piece of kit.
$340 is more than the cost of the rest of the BOM, and it feels kind of crazy that the plastic housing would cost more than the electronics.
I get that the parts are 3D printed, but are they being printed on-demand or something? I wonder if the 3D printing files have been open sourced - would be interesting to hear from a 3D printing enthusiast here about the costs (it's HN, there's bound to be many!).
stl files are linked in a zip at the bottom. I've never touched a 3D printer before so I don't know what you do with them, the online 2d printers I see want me to create an account to generate a quote which seems tedious.
STL files are the 3D models that you put into your slicer of choice (which converts the project into the thin layers that the printer will gradually put down).
There's a dozen things to print in the zip. I sliced 3-4 of them, and none seemed particularly heavy. Something like $0.50 of filament each for what I pay for PLA, which would put the whole thing at ~$6 of material. Perhaps the cost is being driven by some sort of print-on-demand costs?
FWIW, you could also buy an Ender 3 for ~$200, and $140 buys a lot of filament, so you could get into the 3D printing hobby for less than the quoted cost.
You are significantly underestimating the cost of producing molds. You could easily hit $100k for molds made in the USA, $10-20k made in China. That’s also assuming you get the molds right the first time and don’t have to go back and make new ones. Plus there is the setup and minimum production runs, you need a fairly large demand to make producing these viable. 2-3 of these ($680-$1020) doesn’t cover cost just to modify the design to work for injection molding.
To make enough shots for a low unit cost, the mold has to be made of very hard steel. So that takes a long time on a very expensive cnc. If you make it out of softer material to make the mold less expensive (ie aluminum), you only get ~100s shots or so. Good enough for testing / beta runs, but not enough to amortize the engineering & cnc costs.
Prototype tooling can sometimes (depends on your part obviously) get you up to the 1-10k range of shots now. At that range, it can be sufficient for certain types of product runs.
I've worked in the molding industry. First, the mold has to be designed. This involves planning for the flow of molten plastic through the passages and cavity of the mold, venting of air, often re-design of the part so it can be effectively molded, etc. Shrinkage as the part cools affects how multiple parts fit together. That's huge for consumer products. There's CAD software for doing this, but it still requires an expert to know how to use it.
Second, the mold itself is made of a very hard steel, though there are cheaper "soft tooling" materials for making lower quantities before the tool is worn out. The special steel may require special machining processes. The tool also has a variety of moving parts such as ejector pins and other mechanisms if the part is not shaped to be ejected cleanly.
If the mold comes from low-cost supplier, there's the added cost of reworking it in order to iron out minor problems. I spoke to one tool-and-die machinist who told me most of his income came from repairing cheap molds.
If you save on the mold, you often pay more for the pieces due to scrappage, difficulties setting up the molding machine, or the short lifespan of the mold.
Right now I work in a business that makes expensive high-tech equipment, so we can afford to eat the cost of making expensive parts from expensive molds.
For prototype quantities, we tend to prefer things like folded sheet metal, CNC, and even small aluminum castings with secondary machining processes. In that milieu, the cost of 3d printing is actually quite competitive. I just got some pieces costing about 1000 bucks each, but they were delivered to my doorstep a few days after being designed.
Also: Something I've learned is that everything I know might be wrong tomorrow. Every manufacturing process is being steadily improved, and a process that you wrote off as being too expensive might have suddenly become cheap for reasons you're not aware of. Always ask.
Thanks. I find this whole area fascinating. It feels like if there was a way to make molded parts cheaper it would open up a large gulf in between "$300 3d printer" and "$15k for one mold and BTW you need 5 molds for your thing".
I do a lot of 3d printing but have run in to the limits quite a few times where I want something bigger than my print volume allows, or I want something water proof, or food safe, or I want something transparent (which can also be used outside), or I want it smoother...
resin printing or finding someone with a big printer... these things help sometimes.
Another trick I have yet to succeed at is using some existing plastic part, or transparent plastic from a soda bottle... Repurposing for these edge cases.
Yeah I'm not sure about that either... All the parts look FDM printed too which is definitely way cheaper than that if you already own a printer. Perhaps they had some company print the parts?
Honestly for that kind of money you're getting kinda close to having them SLS printed which would be much sturdier and look nicer...
I assume a person or small team is trying to make up for the time spent desiging this?
You can't just add up all the components, add 10% on top and expect a price level that competes with Apple. Looks like a hobby project to me and there is no harm in supporting the creator.
Filament is really cheap. Basic pla is roughly 2-3c a gram. I couldn't see this being more than 500 grams, so maybe $10-20 for filament, maybe $30 if you use abs or something.
I'm guessing the rest of the price is paying to have it printed by a third party.
I've been looking into something like this, and it seems the display is the hardest part. I haven't been able to find a large monochrome LCD that is compatible. Now if you don't care if the screen uses a lot of power, is color, etc. then there are a lot of options similar to what you're describing on reddit.com/r/cyberDeck
You can do it for ~$30-40. There are two problems: The screen, and the microcontroller. For the screen, you can just use a regular TFT, and only display one color. This has the benefit of bringing down the size of the frame buffer down from 800x480x8 or 16 to 800x480x1. That's still a lot of RAM for a microcontroller, but you can get displays with onboard graphics RAM, and any small microcontroller can use the on-board GRAM as a framebuffer. Line scrolling is a bit of a trip (kind of like a bubble sort, but on lines of pixels), and blinking the cursor is _weird_, but yeah, it can be done on a $1 microcontroller and a $5 screen.
Blinking used to be accomplished by XORing the pixels with '1'. The nice thing is it doesn't use any memory to restore it since you just XOR a second time to get what you had back.
> full-size keyboard, without the touch and without the GUI. Just a term. Think tandy model 100.
Yes! With a nice mechanical keyboard. I found a nice display for something like this a while ago; these 1280x400 pixel, 191 mm wide (visible area) IPS displays originally meant for automotive use:
I don't think getting a reflective LCD display in this form factor is realistic these days, they stopped manufacturing those in the 80s. I bought this OLED 256x64 display to play with- it's got enough pixels to make it work, but it's just too small/high density to work well form-factor-wise with a full width tenkeyless keyboard.
The cost is very high for just the materials, even rounded up to the nearest reel. On the other hand, you'd be lucky to buy an ABS-capable printer for under $340, considering they generally require a thermal shroud to print reliably.
I added an rpi0 to my Olivetti M10 (a Model 100 clone) and it is wonderful! So much data, a great keyboard, and a reason to sit and read and stop bothering about pixels .. ;)
A guess (based on my experience with RPi): It’s handy if you can debug your RPi peripherals using a “normal” computer to make sure they work. You can even develop an application on a desktop with a fast CPU (for faster compilation) and then ship it to the RPi when you’re done.
I started to build a "compact" handheld RasPi 3B+, but the driver situation for all the shield-like display assemblies I could find (including the promising-looking one I gambled on buying) was very poor. The best I could find was "download our special build of Raspbian", which is a showstopper for my priorities.
If there shield-like RasPi touchscreen open source drivers, mainlined in Linux, and also included in Raspbian and Debian, I'd like to hear of it.
(Blob-free would also be nice, though the RasPi itself has blob problems, so isn't as good as it could be for open source.)
I get flashbacks to the college intern I interviewed that excitedly told me about the RPi telemetry design he created for a rocket. Got it all coded up and working great.
After some prodding he confessed that it failed on launch. The micro SD connector couldn’t take the vibration.
That's odd - I've flown RPis on drones with far more vibration than the relatively smooth linear acceleration of a model rocket. In such setups it's recommended to directly solder power and USB connections, but I've never heard anything about the microSD connector. Did he say how he diagnosed the failure?
Yep, the connectors on those things can have intermittent contact. If you're uploading data to the SD card live while in flight, those intermittent connections can screw up your data. The workaround is to use a flash chip to write live data to it and then once the rocket lands, dump all of it to the SD card.
Inspired by things like these, I bought a 5" LED display for my RPI. The idea was to display the time / weather on Chromium in a "kiosk mode". I was however surprised that with the latest install of Raspbian OS it was even hard to open a simple webpage on Chromium. It was just too slow. I would really avoid buying / building anything on a RPI from now.
That's Chromium on ARM under pure Linux, not necessarily an isolated RPi issue. I've found both Chromium and Firefox to be unusable on even Jetson with Nvidia graphics but a similar specced fan-less x86-64 Chromebook handles browsing with ease as Google has done a fantastic job in optimising ChromeOS. So browsing experience on ARM SBCs have left more to be desired largely due to lack of memory.
Can anyone share their browsing experience on 8GB RPi?
I'm using Firefox on a 4GB RPi4 daily, without problem (with uMatrix that disables unneeded scripts however). It is a bit slow on media-heavy sites (youtube, ...).
It tried the same software configuration previously on a RPi 3B+ (1GB): there, Firefox was quickly exhausting the RAM.
Yeah, the idea was to just display a nice SVG + CSS + HTML clock with Chromium running in full-screen mode (screensaver off etc.). I know it is not the most efficient option but I was surprised that RPi struggled with opening a HTML page on localhost.
I'm running RPi3's in Kiosk mode with a somewhat complicated UI (realtime graphs and such). It's a bit slower on some actions, but plenty usable.
If you want to try it again sometime and are up for trying Elixir, look into the Nerves based kiosk. It builds an image specifically for the RPi3, and you can run Phoenix LiveView and entire system for less than a hundred megs of ram: https://github.com/nerves-web-kiosk/webengine_kiosk
Fwiw, 2 years ago I tried to develop an app at work on Windows IOT on a Raspberry Pi 3. I abandoned pretty quickly, the thing was having a hard time just scrolling some text.
At the time I thought Windows IOT was not optimized, but maybe it's just the whole x86 ecosystem that hasn't been adapted to ARM computing.
I think you can install Android on a Pi, it would be interesting to see if Chrome is fast enough there.
A veritable steal at only $629.99, truly emphasizing the philosophy of the Raspberry Pi project. I'd seriously consider this in favour of my next real estate purchase
even though it's canadian $ it seems highly overpriced. A raspberry pi of the last generation for almost double the price of the current gen (36$)?
And 3d printing 340$.. did the author mean buying the 3D printer too? Because I'm sure if you have a friend who has a printer they can print it for 10$ and they make a 8$ profit with that.
IMO: the pinephone feels very much like a handheld raspberry pi (It even has some IO on the back) and costs a sixth of this (although personally I’ve already found a couple kernel bugs on mine, on the other hand the original raspberry pi was way buggier so maybe I should be happy I haven’t found more.)
I recently found this project and I'm very excited about this DIY portable Linux computer trend and hope some of them graduate to mainstream. N-O-D-E recently announced one [1] and there is another RPi zero based phone project on Crowd Supply[2].
The reason I want a portable Linux handheld computer is, because the direction in which smartphones are moving towards is frightening to say the least. Smartphones are essentially mobile computers and for many parts of the world the first computing experience. COVID-19 situation has shown how mobile computing has been taken for granted and misery of those who have no access to it[3].
Can we imagine a computer manufacturer selling $2000 computer which will receive security updates for only 3 years if lucky, can't install alternate operating systems or software of our choice? Then why are we providing smartphone manufacturers with such privileges? With that unfair advantage they are dictating mobile computing for the entire world. I'm afraid it's just a matter of time, when this trend scale up and affects overall computing in general.
n-o-d-e seems to have the best designed project out of all of these. Like it looks like something I would actually want to buy. Though it does depend on the price. All of these hand-held devices look really attractive until you see the price. For instance, I love the design of the GDP Pocket, but that price keeps me far away.
For me, the single most attractive aspect of the RPi was booting from SD card, no user data or user programs stored on the "device" (computer). Hence, I can run Linux, and/or BSD, and/or Plan9, etc. I can have a selection of different SD cards, each with a different OS. This is how I built systems for myself before RPi. Boot from USB, everything running in RAM, nothing necessary stored on HDD.
However, when RPi alternatives are "advertised" and discussed on HN, the focus usually is solely on (non-bootloader) specifications and price. The option to run multiple OS is seemingly not a concern for any commenter and not a "target" in the manufacturer's marketing.
The best computers for me have been development boards. The reason for this is simple. In most cases they are less biased toward one OS, namely Linux. I probably would feel differently if Linux was favourite OS and had no desire to run anything else. However, I do desire to run other OS, in addition to Linux. Need to have that choice.
This "bias toward less choice" does not stop at the OS-level. Often I see bias toward certain high-level programming languages. This includes the example above from a few days ago. "WebAssembly and Javascript."
When the RPi was first announced there was mention of "running Linux" and was suggested the device would geared toward Python programming. I feared the worst. Thankfully, things turned out differently.
The funny thing with the RPIs is that whatever you install on it, is running in a hypervisor on https://en.wikipedia.org/wiki/ThreadX which is now owned by Microsoft...
Companies like Microsoft have so much cash, they will constantly be poisoning the well of software freedom destroying the last remnants of privacy, long after their own business models became obsolete or out of fashion. I fully expect Google and Facebook will pose the same issue long into the future. Their models too will see their sun set, but they will still work to have their fingers in our pies (no pun intended).
Yes, that's the downfall on all of these sort of things. Nobody seems to make a good mini keyboard these days. Back in the late 1980s there was the Poqet PC and it had a very nice typable keyboard. I don't know why we can't have something similar in terms of small keyboards, but as far as I can tell, nobody makes a keyboard like that now.
That kind of thing had been my dream device as well, but once I get realistic it gets more and more akin to a random phone with PmOS with keyboard but only as a want item, except PmOS only runs on select(scarce) devices. I think the core of the problem is clunkiness of Android.
e: I checked supported devices list for PmOS “just to be sure” — I was genuinely pleasantly surprised
Regardless of what it costs to make at this price or really any I don't see the appeal. I used to get excited about similar projects but with how good my 2 pine phones are already and the pinetab that is getting delivered Thursday I already have functional sleek linux handheld devices with additional inputs/outputs.
I guess I don’t see the appeal either, mostly because an 800x480 display seems painful.
Getting either a Pine phone or Pine tablet sounds better. Do the Pine phones make sense just as Linux devices, not using them as a phone?
I always travel with a small iPad Pro, and with Mosh, tmux, and a really powerful VPS from hetzner (which is so cheap I always leave it running), I feel I always have a Linux system with me that makes a practical writing and programming platform.
Can the Pine phones be plugged into a large monitor?
I thought about doing something like this. But I eventually went with modularization. I 3D print for my raspberry, another one for my 7' touchscreen and got myself a wireless keyboard / mouse. This way I can use each element separate. For instance, sometimes I use my PI as a server, so I don't need a screen or keyboard. I found myself wanting the screen decoupled from the computer as well, cause it can be used with other devices. Heck, I even played xbox on it once (not so great experience). Here's my setup if anyone's interested:
https://www.youtube.com/watch?v=F88CSWH8SEY
In this video, I'm using the PI as a retro game machine (I can swap the card to boot to other system)
It looks like a solid shell for RPi, it's just a shame that the price is a bit on the high side though.
For an alternative hackable laptop like shell for RPi, consider the CrowPi2 [1]. It does not support touchscreen but the screen estate is much bigger (11.6") and can be configured with many sensors for testing and developing IoT based projects. As an added bonus the price is quite affordable for students and hobbyists. The review of CrowPi2 can be found here [2].
So many people complaining about the price have obviously never tried to build something similar and sell it. Why do you all expect free designs and then complain about big companies being unenthical?
95 comments
[ 6.3 ms ] story [ 475 ms ] threadI bet you could CNC some injection molds the cost of 2-3 of these chassis. If someone could do that and sell the housing for ~$30 each, this would be an impressive piece of kit.
I get that the parts are 3D printed, but are they being printed on-demand or something? I wonder if the 3D printing files have been open sourced - would be interesting to hear from a 3D printing enthusiast here about the costs (it's HN, there's bound to be many!).
There's a dozen things to print in the zip. I sliced 3-4 of them, and none seemed particularly heavy. Something like $0.50 of filament each for what I pay for PLA, which would put the whole thing at ~$6 of material. Perhaps the cost is being driven by some sort of print-on-demand costs?
FWIW, you could also buy an Ender 3 for ~$200, and $140 buys a lot of filament, so you could get into the 3D printing hobby for less than the quoted cost.
Second, the mold itself is made of a very hard steel, though there are cheaper "soft tooling" materials for making lower quantities before the tool is worn out. The special steel may require special machining processes. The tool also has a variety of moving parts such as ejector pins and other mechanisms if the part is not shaped to be ejected cleanly.
If the mold comes from low-cost supplier, there's the added cost of reworking it in order to iron out minor problems. I spoke to one tool-and-die machinist who told me most of his income came from repairing cheap molds.
If you save on the mold, you often pay more for the pieces due to scrappage, difficulties setting up the molding machine, or the short lifespan of the mold.
Right now I work in a business that makes expensive high-tech equipment, so we can afford to eat the cost of making expensive parts from expensive molds.
For prototype quantities, we tend to prefer things like folded sheet metal, CNC, and even small aluminum castings with secondary machining processes. In that milieu, the cost of 3d printing is actually quite competitive. I just got some pieces costing about 1000 bucks each, but they were delivered to my doorstep a few days after being designed.
Also: Something I've learned is that everything I know might be wrong tomorrow. Every manufacturing process is being steadily improved, and a process that you wrote off as being too expensive might have suddenly become cheap for reasons you're not aware of. Always ask.
I do a lot of 3d printing but have run in to the limits quite a few times where I want something bigger than my print volume allows, or I want something water proof, or food safe, or I want something transparent (which can also be used outside), or I want it smoother...
resin printing or finding someone with a big printer... these things help sometimes.
Another trick I have yet to succeed at is using some existing plastic part, or transparent plastic from a soda bottle... Repurposing for these edge cases.
How can the 3D printing be that much?
Honestly for that kind of money you're getting kinda close to having them SLS printed which would be much sturdier and look nicer...
You can't just add up all the components, add 10% on top and expect a price level that competes with Apple. Looks like a hobby project to me and there is no harm in supporting the creator.
> 3D printing, filaments, etc. 340.00
Is this just for the cost of the raw printstock? Or does it include the price of a low-end printer?
I'm guessing the rest of the price is paying to have it printed by a third party.
Anyway, here's what I'm working on: https://twitter.com/MrRobotBadge/status/1274490808616419328
It's based on the VT-100. It's got a 69 key keyboard. The name VT-420 was already taken, so it's called the VT-69.
Yes! With a nice mechanical keyboard. I found a nice display for something like this a while ago; these 1280x400 pixel, 191 mm wide (visible area) IPS displays originally meant for automotive use:
https://www.buydisplay.com/catalogsearch/advanced/result/?re...
I don't think getting a reflective LCD display in this form factor is realistic these days, they stopped manufacturing those in the 80s. I bought this OLED 256x64 display to play with- it's got enough pixels to make it work, but it's just too small/high density to work well form-factor-wise with a full width tenkeyless keyboard.
https://github.com/raspberrypi/linux/issues/2521
its possible, but hard and buggy
If there shield-like RasPi touchscreen open source drivers, mainlined in Linux, and also included in Raspbian and Debian, I'd like to hear of it.
(Blob-free would also be nice, though the RasPi itself has blob problems, so isn't as good as it could be for open source.)
EDIT:
if you want to have a look at the installer script it's here: https://github.com/adafruit/Raspberry-Pi-Installer-Scripts/b... (the business end starts at line 222)
https://github.com/librerpi/rpi-open-firmware https://github.com/itszor/vc4-toolchain/issues/7
I get flashbacks to the college intern I interviewed that excitedly told me about the RPi telemetry design he created for a rocket. Got it all coded up and working great.
After some prodding he confessed that it failed on launch. The micro SD connector couldn’t take the vibration.
Can anyone share their browsing experience on 8GB RPi?
It tried the same software configuration previously on a RPi 3B+ (1GB): there, Firefox was quickly exhausting the RAM.
If you want to try it again sometime and are up for trying Elixir, look into the Nerves based kiosk. It builds an image specifically for the RPi3, and you can run Phoenix LiveView and entire system for less than a hundred megs of ram: https://github.com/nerves-web-kiosk/webengine_kiosk
At the time I thought Windows IOT was not optimized, but maybe it's just the whole x86 ecosystem that hasn't been adapted to ARM computing.
I think you can install Android on a Pi, it would be interesting to see if Chrome is fast enough there.
And 3d printing 340$.. did the author mean buying the 3D printer too? Because I'm sure if you have a friend who has a printer they can print it for 10$ and they make a 8$ profit with that.
The reason I want a portable Linux handheld computer is, because the direction in which smartphones are moving towards is frightening to say the least. Smartphones are essentially mobile computers and for many parts of the world the first computing experience. COVID-19 situation has shown how mobile computing has been taken for granted and misery of those who have no access to it[3].
Can we imagine a computer manufacturer selling $2000 computer which will receive security updates for only 3 years if lucky, can't install alternate operating systems or software of our choice? Then why are we providing smartphone manufacturers with such privileges? With that unfair advantage they are dictating mobile computing for the entire world. I'm afraid it's just a matter of time, when this trend scale up and affects overall computing in general.
[1]https://n-o-d-e.net/zeroterminal3.html
[2]https://www.crowdsupply.com/arsenijs/zerophone
[3]https://needgap.com/problems/149-remote-education-for-underp...
https://davejansen.com/installing-ubuntu-18-10-on-a-gpd-pock...
https://www.aliexpress.com/item/32968851612.html
0: https://images.app.goo.gl/sNcy4QBQwXDXmyVj7
That choice should include more than just "Linux".
Many, nearly all "Raspberry Pi competitors", fail to offer that choice.
Consider this one from a few days ago:
https://news.ycombinator.com/item?id=24378758
For me, the single most attractive aspect of the RPi was booting from SD card, no user data or user programs stored on the "device" (computer). Hence, I can run Linux, and/or BSD, and/or Plan9, etc. I can have a selection of different SD cards, each with a different OS. This is how I built systems for myself before RPi. Boot from USB, everything running in RAM, nothing necessary stored on HDD.
However, when RPi alternatives are "advertised" and discussed on HN, the focus usually is solely on (non-bootloader) specifications and price. The option to run multiple OS is seemingly not a concern for any commenter and not a "target" in the manufacturer's marketing.
The best computers for me have been development boards. The reason for this is simple. In most cases they are less biased toward one OS, namely Linux. I probably would feel differently if Linux was favourite OS and had no desire to run anything else. However, I do desire to run other OS, in addition to Linux. Need to have that choice.
This "bias toward less choice" does not stop at the OS-level. Often I see bias toward certain high-level programming languages. This includes the example above from a few days ago. "WebAssembly and Javascript."
When the RPi was first announced there was mention of "running Linux" and was suggested the device would geared toward Python programming. I feared the worst. Thankfully, things turned out differently.
Companies like Microsoft have so much cash, they will constantly be poisoning the well of software freedom destroying the last remnants of privacy, long after their own business models became obsolete or out of fashion. I fully expect Google and Facebook will pose the same issue long into the future. Their models too will see their sun set, but they will still work to have their fingers in our pies (no pun intended).
https://en.wikipedia.org/wiki/Poqet_PC
A device with a monochrome screen, 800x480 resolution, 3D case with a qwerty keyboard shoudn't exceed $100.
They day a little netbook/palmtop arrives with Linux at a ridiculous price (the one of an RPI), it would be almost a second revolution as the zaurus.
Portable Nethack/IF/SSH/IRC/Usenet + coding, PDF/image/video viewer with framebuffer support, cheap audio from a SOC if any.
That would be my dream device.
http://pyra-handheld.com
https://openpandora.org
Disclaimer: I love my OpenPandora!
e: I checked supported devices list for PmOS “just to be sure” — I was genuinely pleasantly surprised
$629? As someone else stated, ouch.
Getting either a Pine phone or Pine tablet sounds better. Do the Pine phones make sense just as Linux devices, not using them as a phone?
I always travel with a small iPad Pro, and with Mosh, tmux, and a really powerful VPS from hetzner (which is so cheap I always leave it running), I feel I always have a Linux system with me that makes a practical writing and programming platform.
Can the Pine phones be plugged into a large monitor?
Is there no touchscreen / smallscreen friendly Linux desktop shell?
For an alternative hackable laptop like shell for RPi, consider the CrowPi2 [1]. It does not support touchscreen but the screen estate is much bigger (11.6") and can be configured with many sensors for testing and developing IoT based projects. As an added bonus the price is quite affordable for students and hobbyists. The review of CrowPi2 can be found here [2].
[1] http://linuxgizmos.com/hackable-crowpi2-steam-education-lapt...
[2] https://magpi.raspberrypi.org/articles/crowpi2-review-raspbe...