Collapse OS, An OS For When The Unthinkable Happens

Decades of post-apocalyptic Hollywood movies have taught us that once all the trappings of our civilisation have been stripped away, it’s going to be kinda cool. We’re all going to wear slightly dusty looking 1980s motorcycling gear, and we’re going to drive really cool cars. Except of course Mad Max is fantasy, and the reality is likely to be unspeakbly grim. The future [Virgil Dupras] is anticipating is not a post-nuclear wasteland though, instead he’s trying to imagine what access to computing might look like in a world where the global supply chain has broken down. His solution is CollapseOS, an operating system designed for resilience and self-replication, that runs upon the minimal hardware of an 8-bit Z80.

It’s a pretty basic operating system so brace yourself if you are expecting a 64-bit fully multithreading kernel. Instead, you’re looking at a kernel, an assembler, and a text editor. One of the stated aims is that it can compile assembly language for a wide range of target CPUs, but it does not make it clear whether this means the OS itself will support those platforms. The self-replication is a fascinating feature though.

It’s an interesting question: what computing hardware would be available to the would-be hacker in a world in which all parts must be scavenged? The Z80 and other processors like it fit the bill admirably in one sense as it is possible to create a working computer using them with fairly minimal tools and knowledge, but we can’t help wondering whether the days when almost any electronic junk pile would contain one are now past. So what other easily accessible computing platforms might be created from post-apocalyptic junk in 2019? Remember, with no laptop and IDE you can’t just put an Arduino bootloader on that ATmega328 you desoldered from an old thermostat. As always the comments are open.

Image: Damicatz [CC BY 2.5].

180 thoughts on “Collapse OS, An OS For When The Unthinkable Happens

    1. One if the first uses of personal computers for farming was weather and crop yeild tracking along with stock breeding genealogy. Just moving past paper ledgers and the ability to run simple spreadsheets greatly reduced the administrative burden on farmers, allowing them to concentrate on actual farming. So yes, it would help grow potatoes.

        1. Tracking crop yields is not admin, it’s basic farming going back hundreds of years. Knowing which fields have the best performance with what crops allows for better crop rotation to avoid depleting the soil nutrients. Tracking weather patterns year over year, especially in a post-crisis world, will be crucial. Ancient farmers did this on wood slates. Good farming even for a very small community requires good data and long range thinking. Farmers who just stick seeds in the ground and hope for the best every year starve eventually.

        2. Lack of well documented things may look trivial today. But lack and mixup of stories and genealogies have led to current religions. And we were unaware of numerous Kings, cultures and societies and still are , before archaeologists happened

          1. There was a woman on the radio the other day talking about planting trees in a city. She’d planted thousands of trees with her organisation, including apples. I’m pretty sure that’s what happens if you plant apple seeds.

            Bananas are much more feeble. Possibly the king of hopeless food crops. All bananas are cloned, the brown smears you see in a banana are the remnants of what used to be seeds. Almost all bananas in the world are genetically identical. There was a great banana plague in the 1950s, it was almost the end of bananas altogether. In the end, they found a breed that would still grow and wouldn’t succumb. So now, posibly except in elite banana enthusiast circles, all bananas you can buy are Cavendish. Prior to the ’50s they were Gros Michel.

            Apples are much tougher. Pretty sure you don’t know your onions there.

          2. Most apples you can take a seed out of, condition it correctly (might need cold temperatures, forgot) and get it to grow into a tree… however, the fruit that grows on that is gonna be undersized sour crab apple type things, not the golden delicious or whatever you were expecting. The cultivars all must be cloned to get the right apple characteristics for eating, and this can be done by grafting a branch onto generic apple root stock. So if your buddy has a tree with good apples, plant any old apple, grow it on a couple of years, then graft a branch of his tree on it.

          3. @all the apple naysayers: look up “F1 hybrids”. For wannabe independent growers it’s almost like patents, but enforced by physical realities. Thankfully, there are varieties which don’t need to be hybridized to give good apples.

      1. Not just personal; an agribiz was one of the first companies in my old hometown to get a computer. The funny side is they bought it for dreams of a paperless office, but this was in the 60s… :D They had this one huge filing cabinet, and they hated it. They bought the computer thinking they’d be able to get rid of it. They ended up needing two more of these huge filing cabinets for the computer! (Maybe it wasn’t “paperless office” they were thinking, but intermediate computations or something, I don’t really know. I didn’t exist in the 60s.)

    2. Try making paper COMPLETELY from scratch, then ask this question again.
      Also things like controlling irrigation, keeping a callendar, keeping track of yields when trying new breeds, keeping track of what was planted where so you can rotate crops…all that are things that computers can make much easier, leaving you with more time for actually growing said potatoes.

      1. That’s a very interesting point. It could be that the collapse would lead to a paperless society, if computers become easier to obtain than paper. But that would require more than computers – it would require digital communications as well, even if this was mostly by encrypted sneakernet.

        1. Interesting topic to add the deforestation angle where paper wouldn’t be as common due to farming that is not timber or related to timber like sap, distillate, extract or nut farming. Maybe the finer fibers as food sources would be banned or something.

          I’m on the band wagon (or riding solo) that there isn’t going to be a collapse… only cycles like Kondratiev cycle/wave or other time range economic cycles that might be due to 15 year old’s being the largest percentage of the population in a region (cause war), fractional reserve banking with bond markets or other schemes that cause extreme societal ranges for periods of time due to the availability of capital. Maybe extreme weather conditions can cause an issue also… though I’m thinking there are enough ICBM’s and brain trust know how left on the planet that can divert the trajectory of something from outside the planet.

          1. Actually paper is pretty easy to make from hemp. You could make it from marijuana as well but then your potato chip needs would multiply, not to mention avocados for guacamole putting more pressure on the farmers to maximize efficiency.

      1. And an unlimited supply of 7.62 x 39? You could raid and probably wind up getting shot or killing the guy raising the potatoes, or you could become a security force for the farmer and earn a wage in potatoes for keeping his farm safe. Better long term results with the latter plan. Something to consider.

      2. So let’s get this straight… you come onto a maker website to announce your plan is to steal from makers? Farmers are makers you know.

        Nicely done. Hope your first raid on your fellow American is also your last raid.

    3. If we have power. Can use solar and wind, but not with out electronics. The end comes when a solar flare that is so big cooks all the gear made and it doesn’t matter anymore.

        1. Well said! Even for generating electrical power from wind, you don’t need any silicon. The worst case is you use brushes and a split ring in the generator to produce DC, use that to charge batteries, and use them to supply smooth power for your electronics, switching over manually. …But I think even that gear can be cooked by a big enough solar flare. Telegraphs were not any more sophisticated than that, although possibly more delicate.

    4. The need for computers after the collapse of society will be exactly the same as the need for computers prior to society.

      You make a good point, that computers don’t directly solve traditional survival needs. Therefore, please throw away all of your computers since you don’t need them.

    1. The correct Z80 troll answer is “What is a 6502? I have never heard of it!” :)

      It suffers from the same scarcity in junk as the Z80 in 2019, with the added handicap of less appearing in embedded applications. I’ve frequently unearthed 40-pin Z80s on industrial boards, never a 6502.

      But yes, it would be as suitable if you could find one.

      1. Both of these CPU are still in production in their CMOS sibling. Zilog still sell it in its through hole 40 pin PDIP form. Western design center still sell the 65C02 in 40 pin PDIP format too. So one should conclude there is some of them scavenge.

        1. I still have z80’s and 8085’s in my chip box, even some of the old memory chips in there 1x64k dynamic. Lots of old TTL and some CMOS stuff too. Even a Votrex voice chip and the original radio shack voice chip with the added external rom. I knew the Sales manager for Votrex and got some of the chips on eval in 86′ I think. I built multiple versions of different types of voice synthesizers at the time, but they are nothing compared to the text to speech and phonetic stuff software can do now. Back then the DEC Digitalker was really the only one in commercial use if I remember.

          1. Strictly speaking, C was designed as a feature upgrade from B, which itself was BCPL “bikeshedded” by Ken Thompson to suit his taste for minimalism. This minimalism together with the necessity of coding efficiently in that era meant they ended up with a decent language. It was only after C was made that they realised they could implement most of the Unix kernel with it.

            I prefer Forth, myself. It’s a far more powerful language, being extensible. The amazing compactness of Forth code can be highly valuable with the limited memory of 8-bit computers. And, it’s a better fit to those old computers than it is to modern ones with many registers.

            In fact, I’m trying to produce a Forth-based rival to CollapseOS. :D (“Trying” because my health holds me back. :( ) It’s not quite a rival though; I’m targetting BIOS-PC. If, after The End (lol), you find yourself in an industrial area, there should be no shortage of BIOS-PC-compatible industrial computers. Knowing a little about the way industry operates, I think this will be the case for at least another 20 years. From headless process controllers to wall-mounted displays, there’s a lot of BIOS-based stuff out there.

        1. C was not widespread on self-hosted 8-bit OSs. Even today, while 8-bit C complilers exist, they’re so limited they are a fraction of developers. You would be changing floppies multiple times to build your application, or using up to 4 drives. No way (there’s probably more folks doing cross-compile C from a real OS)

          If you’re coding on ye old metal, you do so in Assembler or the ubiquitous BASIC.

          1. That’s right, C was not generally to be found on 8 bit computers. But I believe that one of the things that made microcontrollers popular in the hobby world was Arduino, and the part of Arduino that made that work was the C++ compiler. Nobody wants to try to keep the instruction sets of multiple CPU architectures in their head at the same time.

            As for the practicality of C on 8 bit computers, you seem to be assuming that floppy disks would be the file storage medium. Highly unlikely, since the methods for using SD cards for this are widely known. So it it requires six passes to do a compilation on a limited-RAM machine, you’re not going to even notice, because the SD card doesn’t make any noises, nor does it take much time to transfer a few kilobytes to memory.

            Collapse OS isn’t about finding old computers and putting them to use. It’s about building 8 bit computers using chips found in otherwise useless modern devices and putting them to use, when for various reasons our big, bloated PCs won’t work any more.

      1. This is why I advocate a multi-architecture approach, distributing simple applications in source code in a high-level language. So it DOESN’T MATTER if the hardware you end up with is 8080, or 8085, or Z-80, or 6800, or 6502, or 6809. If you want to know what will work after the Internet, look at what worked BEFORE the Internet. People published magazines full of BASIC programs that could be run either as-is or with minor modifications, on any of the home computers of the time.

    2. Yes. The Z-80 was never the dominant MPU, because the MPU divergence started with the split between the 8080 and the 6800, so there never WAS a universal MPU. A useful OS for post-collapse must have source code available and pre-compiled kernels for a variety of architectures, and a C compiler that can make it architecture-independent. I’m not suggesting a Linux-like thing; OS-9 already does that, but it needs a CPU at least as good as a 6809, which isn’t a given when you’re scrounging for MPUs. I’m thinking something more basic like a CP/M for multiple architectures that is source-compatible across architectures. So the device drivers for built-in hardware and a very basic core library (less comprehensive than GNU libc, for example) for a given machine might be done in the native machine language, but applications would be distributed as source code, and the installation would involve compiling on the user’s machine. Sort of the way Linux used to work, before builds got too complicated to “just work” on any Linux machine. But that level of complication probably wouldn’t happen when machines are expected to be limited to 64 KB of memory, and as long as the OS shipped with the compiler, no problems with different flavors of C.

      Remember that CP/M and other 8-bit OSs were limited not only by the 64 KB memory space, but also by the capacity of floppy disks. A modern replacement would be geared toward SD cards instead, so throwing in things like a compiler and a PDF viewer wouldn’t be a major problem.

  1. “Remember, with no laptop and IDE you can’t just put an Arduino bootloader on that ATmega328 you desoldered from an old thermostat.”
    I think this is aimed more at a case where “there is at least enough around in order to restart the tech industry quickly”, as was mentioned in the “why” page for the project.
    If society shits the bed too hard, there’s no easy coming back, it’d be decades before we got everything retooled and working.
    If it is just right, however, there will still be plenty of working electronics around to create new hardware that we can build easily without massive, technically impressive factories. Including laptops to slap some OS on [insert SBC] or whatever else.
    Simple electronics are easy to fix and connect together. They also last a very long time, while more complex components fail much easier due to their tiny nature. You can even make systems that could run this OS using basic components, sans some embedded IC platform, like the good ol’ days.

    As Kahoover also mentions, this is more the thing that would be done after you were able to get reliable food sources in order too, which would only really be doable in a not-totally-doomsday scenario where anarchy rules the lands Mad Max style.

    1. The thing is, since the ATMega328 and many similar chips are programmed by SPI and the chips don’t have a lower limit on the SPI clock frequency, you can actually bit-bang a program in by hand.

      I recently stumbled on a tutorial to load up a simple blink program literally by toggling switches on an ATTiny, but I can’t find the page any longer.

      So yes, while you can’t put an Arduino bootloader, armed with the documentation for the chip, you can program them and do a whole lot more with that ATMega chip you pulled out of a discarded smart lightbulb.

        1. “Without a computer”? What kind of question is that? Both the original chip and the target ARE computers. I mean, of course it would be possible to design a circuit to generate the waveforms that would read words from one chip and write them to another, but I’m not sure what the point of that would be, since it would be so much easier with even A MICROCONTROLLER to do the job.

      1. It depends on what you mean by “raspberry pi”. If you mean a credit-card-sized development board, then there are dozens of different ones out there. But if you mean a credit-card-sized machine that can run a multitasking OS complete with a GUI system, that’s just not likely to happen. If you mean something like a TRS-80 Model 1 or ColorComputer, they’ve already been done.

        1. There are multitasking GUI OSs for Commodore 64 and Atari 130XE. :) There must be some for Z80 computers. They require some really clever tricks to work, though, and given that those old machines relied on DMA to feed the screen, somewhere around half the CPU’s time is spent idle while graphics controller DMAs the 8KB data for the graphical display. ;) Oh I wonder if that even applies to Atari 130XE if the display uses a separate memory bank… I don’t know. But I’m just freewheeling, I don’t even get on very well with full GUIs.

          Anyway, see “my” website (click my name, I guess,) for a video of an unmodified Atari 130XE running a full GUI, remaining smooth with many windows. Or search Youtube for “Atari 8-bit GUI: Low-level GFX Rewrite”.

    2. I think there are plenty of people working on fabbing chips at the home shop level. Now: LCDs! Don’t need to be color, don’t need to be high-res. Just being able to do something on the order of 8 rows of 40 columns of 8×6 character cells, or the graphical equivalent, 240x 64 pixels, gets us up to laptops, at least at the TRS-80 model 100 level. Which was good enough to do any essential computing.

  2. The CP/M operating system from the 1970’s already came pretty close to meeting the goals of the “Collapse OS”. It ran on the 8080, Z80, and similar CPUs. It came with complete manuals on its inner workings, and a full set of tools to move it between computers. The XLT86 program would even translate it to x86 CPUs.

    The key point in preparing for the apocalypse is EDUCATION! You have to know HOW stuff works to be able to fix, improve, or build more.

      1. CP/M’s BIOS routines were pretty ugly, iirc. Could definitely stand to see a modernized 8 bit os.

        I mean, given that this is thing is just justification for a fun little lark, a better argument could probably be made that the constraints could already be handled in one of the forth flavors. But then, not everyone appreciates programing backwards, and in heels.

    1. Correct, so don’t throw away paper books on gardening, anatomy, math, physics etc. I scanned almost my complete library of technical papers and books but without electricity or a laptop there is no way to read them.
      Most preppers and wannabe survivalist forget this. (I am not but I have a large physical library because I like books)

      1. Buy yourself a spare unlocked phone. Load all of your books onto this phone. Let the phones battery died once you have digitized everything you wish to maintain. Build a small faraday cage for this phone and put a solar charger and a few cables in with said phone and panels and bury the box where it can’t get damaged.

        You’ve successfully archived and stored everything you have in print onto a portable and efficient device.

        You’re now apocalypse proof.

        1. I’ve been doing this for years. Like, since before smart phones. It takes eternal vigilance, because you have to keep moving all of that essential data to media that will still be readable later on. At this point, it’s on microSD cards, with backups on multiple hard drives. Ten years from now it will be something else, and by then I’ll have everything transferred to that format with a protected device to play it back. And there are many thousands like myself. We’re never going to lose it all.

  3. I dunno. I’ve got P4s crammed in every nook and cranny around here. A bit power hungry for post apocalypse use, but as long as you have a bodged heat exchange mechanism you can heat your wasteland hovel/bunker with it too. maybe even some light cooking.
    So I’m envisioning a reversion to 2005 or so. Not bad, I can live with a mixture of Win2K, XP, and some Linux, after all I do have some hardware that wasn’t supported past Win9x as well. (Really, won’t be that archaic, most of them are already running Win7, anyway.)
    As long as I don’t loose my dusty archive of driver and O/S CDs (and floppies!), I should be able to do a clean install of DOS 3.1, Basic, or any Windows from 95 up on whatever archaic hardware that might be found. Hell, I’ve even got a good suite of CP/M software if I’m reduced to digging out the Kaypro.
    So, I’m not really seeing the Collapse O/S scenario. There are obsoleted business PCs stored everywhere. The important necessities are going to be power and communications.
    Oh, and probably the fine art of making your own capacitors.

    1. Are you sure you still have them? Most diskettes I’ve had die after 10 years in the drawer when they lose their magnetic properties. Even CD’s eventually breakdown when the plastic looses the plasticizer in it. They look fine until you spin them up. I worked as far back as on 8 inch Disks, 5 1/4, etc. They all have distinct problems in a few years. And most diskettes will go CRC error and refuse to allow access after a bit. I even rolled my own with the CPM80 that could be loaded on just about anything. DOS is a bit more of a problem since it uses interrupt controllers for the devices on 8088 and above. And yes – I’ve thrown out Kaypro’s, Telcon Zorba’s and Osbornes before, the Compaq SLT was my favorite. Even some computers you have never heard of likely. Actually Eagle made a really good CPM system, it had the beginnings of graphics in the Zbasic on it. There also was MPM, Concurrent CPM/MPM, etc. And of course Apple II (Darth Vader no less – Made by Bell and Howell) with a single disk drive ($500 bucks in 1979, plus 2500 for the computer)

      1. Nearly all my early 80’s Kaypro and DOS diskettes still read just fine 35-ish years later. Same with hard drives of that era – lots of mechanical problems but few magnetic. I’m certainly not trying to refute your experience, just curious why it would be so different. I didn’t do anything special to preserve mine – I live in a dry climate, but can’t think of any other relevant variables.

        On the other hand, most of my CD-R’s from the late 90’s and 00’s are toast, so whatever I’m doing different clearly doesn’t help with those.

        1. “On the other hand, most of my CD-R’s from the late 90’s and 00’s are toast, so whatever I’m doing different clearly doesn’t help with those.”

          Built-in defense against country-western making a comeback. :-D

        2. I suppose magnetic fields – even ones that blanket a neighborhood might have something to do with it. For example: A nearby high tension electric line. It’s possible that the older floppies were built better that we thought. I converted my 5 1/4s to 1.4 disks in the early 90’s after they came out, so I don’t have those anymore. But I remember that many won’t read, and those were 10 years old. I did have some 8 inch disks that read after 12 years, they are lower density and don’t pack everything so close, maybe that makes a difference on the recovery of data reads. My 1.4 disks I converted to hard drive repository and created images of each disk with a corresponding database entry. You can mount the disk image as a drive if needed using some basic software like Magic ISO I think. It’s been a while since I accessed them. I do however keep a bootable 6.22 DOS image and work on software in assembly on there occasionally. Biggest problem with that is things like COM ports and LPT ports don’t map to USB because it didn’t exist back then. There is someone that has a product for sale that will map them though, I just don’t feel like paying bucks for that.

          1. I’m banking on how they’re physically stored. I had boxes of 3.5″ floppy disks. Both in those silly plastic floppy disk boxes and paper boxes the blank ones came in.

            When I stored them, I didn’t discriminate what disks went into what boxes. So things like manufactured floppies such as software I bought at the store and floppies where I saved erm…. “documents” just went into whatever box I was using at the time.

            Sometime in 2005 or so, I found all of my old disks, bought an external floppy and attempted to image them all to a HDD. I discovered that every single disk stored in a paper box couldn’t be read. Every. Single. One. It didn’t matter if it was an original game disk or a disk with school work. It was all gone. But nearly every disk stored in a box specifically made for floppies were readable with the only exception being the AOL disks.

            Not even sure why I kept any AOL disks.

          2. Actually, that speaks volumes. It is possible that the the good boxes blocked magnetic fields – even at a minimal amount would help over time. I’m not sure on the ability of plastic to block low field values. Likely better that a single layer cardboard box though.

          3. The paper sleeves are acidic from the residues left from the paper manufacture. This is the same issue as with early CDs, because with humidity the acidic environment slowly degrades the metal particles on the floppies or the metallized layer of the CD.

        1. I’ve read Apple II and 390K Morrow MD3 disks and COCO tapes of a similar vintage no problem though it seems every other one of my 1.44mb disks often have errors.
          On CDs is it’s originally pressed and not scratched it seems to work but a CDR is hit and miss.
          On flash I have pulled files off an old Sony memory stick and Smart media card but those are SLC flash the more modern MLC flash hasn’t been around long enough.

  4. “Except of course Mad Max is fantasy, and the reality is likely to be unspeakably grim. ”

    Modern submarine is big enough for an average family, once you take out the missiles.

      1. You might need to improvise the tools to do it, but you might be able to take the warheads from the nukes to use as fuel for the reactor. You don’t need a lot of brand new fuel to reprocess rods into something that would work… this is assuming you’re a nuclear scientist or at least really handy with a screwdriver.

        If you did it, you’d have energy for the whole village for a good few decades. The stuff in missiles is a lot higher in fissile isotopes than a reactor needs. The basic principle of working with nuclear stuff is not to get near it. Then don’t get too much of it in one place, just in case.

        Nuclear subs have onboard water and oxygen production, as well as electricity. They have workshops onboard. Onboard sewage processing, satellite comms, really not much they can’t do. Even if you’re enemies with a fairly decent navy, you’re likely to be able to escape or kill them.

        Of course, first you need to find somebody who has a nuclear sub in working condition, but somehow isn’t able to defend it from being stolen by you.

        Fuck it, I’ll just die, much simpler.

  5. I think this is more an OS for OS designs you could easily(relatively here) redesign from documentation, photomask and mass fab using visible light and minimal clean room operations. Much of what we do today we did 30-40 years ago and now especially with hindsight we could be running 3d printers and other industrial control, a crude usenet and fidonet text internet, and a packet radio bridging and mesh network. Of course without the rich and taxing 3d graphics and horrible web 2.0 experience we get now.

      1. Yeah, but that would be useless since you wouldn’t have the hardware to run proper CAD/CAM software to design the parts that you’re about to print.

        You’d be reduced to defining the geometry by script and CSG, which is so time consuming and difficult that you’d sooner just make the part by hand and then make a cast of it.

        1. Nonsense. Applications grow to fit the space available. CAD isn’t that difficult – we were doing it when those machines were the personal workstations of their time.

    1. Yes, many things would evolve differently, with the knowledge already at hand. Look at communications: in some countries, wired telephone networks just never happened, because by the time they got electrified, it was easier to put in cell service. That would probably happen everywhere.

      1. But then, in those very same countries, the demand for higher throughput and lower latency forces them to install fiber-optic cables and an electric grid side-by-side to support the repeaters and routers, so in essence they’re building a “wired telephone network” after the fact. After all, the cellular network itself has a wired backhaul anyways and it’s just the last miles that are wireless.

        1. Also, in those countries which skipped the POTS networks: it actually did happen at one point, but then when the colonials left the local warlords rolled up all the cable and sold it for scrap copper.

  6. Linux with GPG for openPGP formatted messages might make a decent system capable of pony-express-ification. Think a system where couriers compile all mail from a community and sneakernet it to other nodes.

  7. I think in case of a collapse, the only thing I would like to use a computer is for driving a packet radio, and mathematical calculations.

    Communication will be highest priority then, if you’re able to use a low power packet radio, like LoRa to talk to other survivors, your chances of survival would be much better

    1. Better off with a crank generator a 6C4 tube and morse code. As long as you have a diode for the receiver and use another tube for the receiver. The z80 would be the last thing on my list of survival items. Perhaps it would make a decent weight for a worm on a hook.

      1. I can only hope you’re kidding. Vacuum tubes? Maybe you missed the last half-century, when other devices were developed that did a better job with better efficiency. I guess you’re going to have bigger muscles than I will.

  8. Most computers are used in an electrically wasteful manner for make work. I can’t think of many things that demand their operation in a carbon neutral society. Managing economies and taxation has been done with clay tablets by the ancient sumerians. It’s true that they are necessary for calculating orbits for space flight, so they may be kept around if that capability is maintained. If the climate emergency is to be headed off, then all non-necessary power usage needs to be stopped. Computers are certainly useful for managing industrial systems but if industry is the problem, then they are a part of it.
    Radio transceivers I suspect will be built due to their use in war but that will likely be the apex of technology. The custom of building machines with machines will likely die so that will rule out integrated circuits. everything will be hand made by necessity.

    1. No, not all power, just wasted power. This article is about using 8-bit computers, which using the modern CMOS versions, use far less than current desktop computers. Once you factor out idiocy like cryptocurrencies, the amount of power people use for computing is vanishingly small, compared with home heating and transportation.

      Desktop computers are already becoming dinosaurs, and even laptops are losing ground to “mobile devices”, including both smartphones and tablet computers, both of which take far less power than traditional (read: Intel-based) computers.

      This really won’t be a factor after the collapse, because the desktop computers just will never be powered up again.

      1. On the contrary. With the production lines broken down, and all the “mobile devices” constructed so they cannot be repaired, the first thing people do is start digging trash heaps and recycling centers for old but functioning x86 hardware.

        A working cracked copy of windows XP SP3 will be worth hard currency because there’s going to be so many driver discs around for that era of machines and hardware, and software was still distributed completely on CDs.

        For the current generation stuff that relies on online distribution and licensing, things simply won’t work any longer. All your fancy Android pads are useless when you can’t connect to Google. The vast majority of people aren’t going to futz up with Linux or trying to hack a lightbulb to run as a wi-fi router – they’ll dust off grampa’s old Pentium IIIs and load them up with copies of Word 97 and continue doing business.

    2. This wont happen. Any country reducing their power significantly will eventually lose to those that aren’t. You cant fight a war on solar, to put it simply.

      It’s pretty much nuclear power or extinction at this point. Reducing electricity isn’t even in the realm of serious discussion

  9. Forget the chips, unless the zombies can eat plastic there’s always going to be a way to scavenge bits and parts in the nooks and crannies of now empty meatspace.

    The real loss is going to be the databooks. Once upon a time I had a huge library of them, now they’re all .PDF’s on my various machines. True, they probably get regularly archived with the rest of my data, but a big pile of DVD’s full of my old projects isn’t going to help all that much when I need to remember the pinouts of an ’00 and a ’74 so I can reboot the universe.

    In fact, the loss of click-ready knowledge is going to hurt a lot more than the loss of actual clicks. Once upon a time every house had a garden and a copy of “Poor Richard’s Almanac” and at least a clue how to grow food. In 2019 my skills are more along the line of knowing how many nanoseconds a 6″ trace on FR4 costs. Useful in its own context, but not really very filling.

  10. Also depends somewhat on the nature of the apocalypse. If it’s environmental or biological, a lot of electronics might survive. if it’s nuclear driven EMP devices, a lot of electronics will be fried.

        1. Just No. you fall into the danger of being “Technically correct” without being actually correct. The Danger EMP causes is because the grid acts like a giant antenna gathering lots and lots of volts to zap connected devices. Yes, a chip on a shelf *might* pick up a few nano volts (and it might not, the wavelength of an EMP blast is very high) but it’s not enough to fry it on its own.

          1. >A HEMP event includes three waveforms: E1, E2, and E3. The E1 waveform is a fast
            (nanosecond rise time, hundreds of nanoseconds duration), broad-band pulse that disrupts
            systems in general, including long-line electrical systems, computers, sensors, and electronicbased control systems.

            Nanosecond rise time is NOT long wavelength. Fast slew rate EM can induce high voltages. Think metal arcing in microwave. It doesn’t take a whole lot to fry chips.

            You are thinking of the E3 pulse which last a long time with disruption similar to solar geomagnetic effects. It is the type that affect the grid.

            >The E1 pulse near ground level at moderately high latitudes may reach peaks of about 50,000 volts per meter!

            >The E1 pulse of a nuclear EMP weapon is what will do the most damage. While it is relatively easy to protect against E2 and E3 (‘easy’ is a relative term), it is quite difficult (or impossible) to effectively protect against E1 for functioning electrical systems.

          3. >>The E1 pulse near ground level at moderately high latitudes may reach peaks of about 50,000 volts per meter!

            Yes, in the same sense as how in a thunderstorm the field potentials can reach tens of thousands of volts per meter – yet the actual amount of charge that can be liberated by that field potential is minimal and a conducting object perpendicular to this electric field would feel close to zero-volts of potential difference because whatever small charge there is will be instantly equalized.

            This is the reason why the EMP will affect things that are connected to the electric grid, but not your cellphone – there’s not enough free charges around to cause significant electric currents to happen despite the high field potential.

            In everyday life, just shuffling your feet across the carpet can charge you up to that 50,000 volts, but can you measure it with a multimeter? Try. See what happens.

          4. >Think metal arcing in microwave.

            Metal arcing in a microwave is caused by forming a conductive loop that picks up energy from the field. Simply placing metal in a microwave doesn’t actually cause it to arc, and many microwave user manuals actually recommend you to place a spoon in a cup of water, or cover parts of a chicken leg with aluminum foil to prevent it from burning.

            Modern microwave manuals tend to prohibit metal items, because that overloads the magnetron which is build with marginal or inadequate cooling to save on cost (most microwaving takes 1-2 minutes at most).


            “EMP/HEMP filter technologies designed to mitigate EMI threats”

            >Typically, the E1 pulse is too brief for traditional lightning protection to be effective, and only transient protection that can respond to extremely fast rise-time pulses is capable of mitigating E1 EMP. The E1 pulse has a typical rise time of 20 nanoseconds, and reaches 50% of its peak value in 500 nanoseconds, and by the IEC’s definition, completely occurs within 1 microsecond. The critical frequency range for the E1 pulse is between 1 MHz and 300 MHz. The electric field strength of the E1 pulse peaks at roughly 50,000V per meter, with peak power densities reaching 6.6 MW per square meter. Depending on a variety of factors, the peak current induced into an electrical system from an E1 pulse could reach 2500A. The E1 pulse is the most dangerous of the HEMP pulses toward smaller electronics and electrical system not connected to long conductor lines, still damaging integrated circuits.

            Is EDM good enough reference for you? They say the same thing here. At 1200 miles radius, they still shows 25kV/M in their graph.

            The last sentence proved my point. Where are your references?
            Where are your calculations?

  11. I once read an interesting story about society in Australia after a major apocalyptic event. They didn’t have electronics because every source of electricity strong enough to be useful was destroyed from the orbit. So they relied on wind power, but they did have a computer. One constructed from many slaves, each one operating a mechanical panel, all panels were connected by complex system of wires that transferred information mechanically. The slaves were named after functions they performed: adders, multipliers, dividers, and some talented ones performed more complex functions. Quite an interesting concept, won’t you say?

    I believe that in case of end of the world an OS for some microprocessors won’t be as important as having information on basics, like growing food, getting clean water, making medicines, etc. In “The Stand” by Stephen King there is a nice reminder of the value of knowledge – one of characters has appendicitis, his friends decided to attempt a surgery even though they didn’t have skills or enough knowledge. They had tools and medicines from an abandoned hospital. It didn’t work out and they killed him. My point is this: in case of collapse prepare an almanac of every scrap of information people might need after the end, covering wide range of topics over multiple tomes, for example:
    – Bear Grylls cookbook or all you can eat (if you are desperate enough);
    – growing food for dummies;
    – pottery for fun and profit;
    – homes, mines, bunkers and everything else, design and construction;
    – glasswork and glassware, including equipment for chemistry;
    – how to make various chemical compounds and medicines;
    – blacksmithing, metallurgy;
    – making machines to make other machines,
    – maths without calculators and DIY maths aids;
    – electricity from scratch, including design and manufacture of basic motors and generators;
    – field medicine, barn medicine and kitchen table medicine;
    – diagnostics without imaging equipment;
    – surgery with cutlery, or how to help someone, when your hospital was destroyed;
    All books printed with long-lasting ink on paper made from hemp fibers so it won’t self-destruct in next 50 years…

      1. Why not? Because relays are power-hungry and EXPENSIVE, with a single gate requiring on the order of hundreds of milliwatts to operate and multiple dollars to acquire, or some combination of hours and materials to build. Better to go scavenging for chips or even transistors in all of the electronics left behind. We’re not talking about waking up one day to find that technology never existed, but waking up one day to severely reduced supplies of electrical power, running water, food, and liquid fuel, but with most existing hardware still viable. If you can find DC motors, you can make electricity from wind and if you’re lucky, flowing water. If you happen to have had solar panels, you probably never lost all your power at all, but only lost much of your capacity. I won’t go into food – that’s the big problem, but not one the article addresses.

        Most people don’t even think about this, but the greatest recent advances in electronics have been in the reduction of power required, which was absolutely essential for mobile electronics, and will be important in the post- days. (Post-whatever.) So the focus for survival (beyond food) won’t be in digging up old technologies, but recovering and putting to new uses the most modern devices that can be scrounged. Sure, your smartphones don’t have anything they can connect to any more … or DO they? They’re all WiFi equipped, so they don’t need a lot of infrastructure to still be useful.

        I’m still all for making use of Z-80 and 6500 architecture, because both of these have been made in CMOS for decades, and while not as much effort has gone into getting their power requirements down, compared to ARM-based MPUs, leaving them at an energy-per-instruction disadvantage, they’re still orders of magnitude better than 1970s technology, and several more orders of magnitude better than relays. These and modern microcontrollers will still be useful because the smartphones will be precious commodities, so people will be mining all kinds of consumer crap for their microcontrollers, for the more mundane tasks.

        1. Assuming that the world end will be caused by nuclear warfare or a solar flare big enough to create global EMP, then most of electronics, including chips in your parts bin, will be fried. EMP will cause ESD in any conductor, including leads and binding wires in transistors, ICs, on PCBs of most of equipment, etc. High voltages and currents induced in electrical grid will kill everything that is plugged in at the moment of the event.There will be no power, no communications, no Internet. No cars either, because they are packed with electronics and miles of unshielded wiring. Better stockpile on those relays and thermionic valves…

          1. I love how the prepper sites say buy a late model 60’s car and you will be OK. Last I checked voltage regulators were first transistorized in 1964. So you would need a vehicle that has a mechanical regulator and mechanical points for timing. And I’m still not convinced an E3 wouldn’t melt the windings in things like starters or rheostat based devices – depending on how close you are to the event.

          2. Preppers just ain’t prepared enough. For example first things to go after major apocalyptic event will be all fuels on gas stations. All streets and roads will be clogged with broken cars or car wrecks. Another problem: if you store your worst case scenario car for 10-20 years waiting for disaster, there is a good chance it won’t work. And you won’t be able to order spare parts. Motorcycle would be better. But not all that gasoline rubbish. Just get a diesel and use any oil after some processing. Or two-stroke motorbike that can be coaxed to run on other fuels than gasoline. For power generators I would consider hot-bulb engines as they look like something that would work for ages on anything that is slightly combustible…

    1. +1

      Dead Tree format. Forget Kindle and PDF’s.

      Most of those books are available on Amazon.

      Here’s a few:

      Caveman Chemistry
      Where there is no doctor
      Wilderness and survival medicine
      Special Operations Forces Medical Handbook
      Dave Ginegery Series on making your own machines from scrap.
      Audel’s books on machining and repair
      Voice of the Crystal
      Impoverished Radio experimenter
      Books by Paul Hasluck
      Books dealing with off-grid living.

      Basically you want(with a few exceptions) books written early in the last to the mid 50’s.

  12. I may lack sufficient imagination, but I have a hard time imagining a post-apocalyptic world where I would want computing capability if things had collapsed so hard that I would want something like this. It would make doing arithmetic easier but really nothing a good slide rule couldn’t cover well enough. If you are interested in technology, and you are even remotely a prepper, you really should have a slide rule.

      1. The Apollo missions were done with 4 bit computers, and if I recall – some of the original shuttle computers were still 4 bit before they upgraded the systems. Seems to me someone just refurbished a flight computer in the last year on Hackaday

  13. As cool as the old stuff is, I would probably target something more modern. There are landfills full of old x86s and ARMs – think of how many phones get chucked a year. There have also been 17 million raspberry pis produced. These ARMs are more likely to run on solar power (we sure as heck aren’t going to have a working power grid) and access massive troves of human knowledge from USB sticks.

    Maybe I should throw a pi and some sticks loaded with my favorite TV shows into an ammo can for later :)

    1. Rope memory reduces the number of diodes you need and can be made quite durable.

      I rather like the metal or glass microfiche approach these guys came up with,

      It can be made to start human readable size and inch it’s way down in size with instructions on how to maintain, repair, or build a microscope to read the rest.

      Even if a few of their ideas are a little outlandish they are definitely thinking of interesting solutions to the problem or long term data storage.

    2. Some IBM product, I think an early mainframe, had capacitive ROM. A grid of contacts, horizontal on one side, vertical on the other, and at the intersections, some cardboard or the like. Punching holes in the card would alter the capacitance between the two intersecting points. So you could replace the ROM contents just by punching holes. The mechanism is just a load of strips on a PCB. The reader circuit needs a bit of sophistication to be able to turn a grid of holes into data bits, but overall there’s a lot to say for it.

      The advantage is that cardboard is much cheaper than diodes, and that it scales up in size pretty cheaply and easily too.

      1. That also sounds relatively easy to gin up with some stripboard. You’d probably want some buffers as amplifier quasi analog mode reading columns and rows or something. Wonder how tight spacing you could go, visible but high density… maybe laser print black paper with white holes, toss some weak acid on it and expose it to sun or UV until they fall out. 3D print them too.

  14. I have to say I must agree and keep around a few books, slide rules, and mechanical adding machines, ah can’t forget a typewriter. I figure once the EMP hits the electronics will be all but dead weight. Print out a photo album as not to miss out on those Kodak moments.

  15. I think the main thing that’s missing is a higher-level compiler, like C. The CollapseOS page mentions “compiling” for different CPUs, but I think for it to be viable when people hare mining for 8-bit CPUs in consumer devices, it needs to have a C compiler, so that applications can be distributed as source code, to be used on whatever system you can hack together. Once you have a kernel and a C compiler that can build both the kernel and the compiler itself (as most can), you have the basis for a hardware-independent 8-bit ecosystem. And THAT’s what is required for a post-collapse world.

    1. On their roadmap page, the authors say that a C compiler is out of the question, due to memory limitations. Well, if C is out of the question, then another compiler is necessary. Don’t tell me it can’t be done, because people are MAKING compilers that fit into that kind of space, and in fact there were C compilers for CP/M, which had the same memory limitations. Sure, they required loading different modules from floppy to complete the compile, but this would be much faster using SD based mass storage.

      1. Author here. There are some C compiler on CP/M that are very tight on resources, true, but they don’t come close to Collapse OS’ achievement that is to self-assemble on 8K of ROM and 8K of RAM.

        1. You’re judging by today’s standards. In the 1980s, there were editor/assemblers that fit within 8 KB, no problem. Certainly for CP/M, and certainly they could assemble their own source code, because that’s what they were DEVELOPED on.

        2. I don’t want to denigrate anything you’ve done – this looks like great work. And I can understand (because this applies to me as well) that for someone without the background in compiler design creating a C compiler – especially one with multiple target architectures – would be daunting. But please don’t say “out of the question”. Instead, reach out to people who HAVE written compilers. It doesn’t have to be C; it can be something even lower-level than C. The main thing is that it MUST be retargetable so that we’re not limited to a single architecture that was only used in a few MPUs. GCC did this in a big way, but that’s a project that started with the basic minimum systems that Linux could run in, which required memory management hardware. I’m talking about doing this in a SMALL way.

          Look at LCC. This takes about 20,000 lines of source, so something like it should be fittable into a 64 KB space, and it is fully documented in a book by its authors, C. W. Fraser and D. R. Hanson, “A Retargetable C Compiler: Design and Implementation.”

          1. Please, read my message. I’m not talking about 64K, I’m talking about 8K. BDS C, for example, if a fine C compiler that is very tight on resources. However, it doesn’t self-compile on 8K of RAM.

            This doesn’t mean that Collapse OS shouldn’t have a C compiler. It would be a good addition. It’s just that Collapse OS, to be self-hosting on very tight resources, needs to be written in assembler.

          2. Don’t misinterpret my meaning: I don’t think that a C compiler wouldn’t be a fine addition. It would. However, to be self-assembling on 8K of RAM, it can’t be written in C. It needs to be written in assembler.

          3. “Can’t”? Them’s fightin’ words. :-)

            Does the Collapse computer has mass storage? If it does (even if it’s a teletype with paper tape punch/reader or a pair of cassette recorders), it can still assemble or compile just about any program that fits on tape. It may require multiple passes, lots of swapping, and a lot of time, but it can be done. After all, that’s how it *had* to be done in the early days, when computers simply didn’t have the resources we take for granted today.

  16. Maybe I am missing something, but cpu is a cool thingy, useless without data to work with though.
    And data needs a hardware to live on.
    Imagine getting everything you need for z80 machine and a blank UV EEPROM. Ooops, not going to work.

    1. That was actually the impetus for this project in the first place. The supply of microcontrollers for doing useful things would last for a long time, both from new stock and chips salvaged from manufactured devices. But the missing link, once the PCs are all inoperable, is a way of programming those microcontrollers. The whole point of the project is to fill that gap.

      When I built my first computer, which happened to be based on a Z-80A MPU, personal computers did not yet exist. There were mainframes and there were minicomputers, but both of these were out of my budget. But it only took me a week to design and build an EPROM programmer, based on the programming specifications given in the datasheet for the 2716. Building the eraser was even easier, involving calling around to find one supplier that had a 20W germicidal UV tube, and another that had a ballast for it.

      The collapse will not be (quite) as paralyzing as many people think, because they don’t remember how we did things before we had smart phones and personal computers. The thing that amazes me most, is that we could even find datasheets. But as I recall, the magazine industry was the source of all information, both in their articles and more importantly, the ads from the manufacturers.

      My point is, we did this before, and we’ll figure out how to do it again.

    1. Many, many post-collapse communities will, indeed, not have the chance to have that person around. To them, Collapse OS will be useless. Collapse OS is for the lucky bastards.

  17. You seem to have forgotten about HAM radio operaters. I’ve been to a HAM swap meet and not only do they have tons of the requisite electronics, but most of them could make their radios from scratch. Honestly, radio and computing go together really well.

  18. I have a propeller spin compiler and the arduino IDE on a few Raspberry Pi boards. The tricky bit will be finding a working display… The good thing is if I find a TV with composite input I’m ok.

  19. I’ve been thinking of this OS lately. It has since branched out by focusing on FORTH to become CPU agnostic, so Z80, 8086, 6809 and 6502 are supported. They also have a Dawn OS to use where stuff is still mostly working. Anyway, I think we can add AI as another possible reason where we might need these OSes. Especially considering we’re close to having AI write AI, so no futzing with learning and testing corpuses, or defining success criteria. “Maximize the health of the population.”… oops.

Leave a Reply

Please be kind and respectful to help make the comments section excellent. (Comment Policy)

This site uses Akismet to reduce spam. Learn how your comment data is processed.