Dusk OS: 32-bit Forth OS. Useful during first stage of civilizational collapse | Hacker News Reader

Dusk OS: 32-bit Forth OS. Useful during first stage of civilizational collapse

duskos.org

169 pointsthunderbong141 comments

> Dusk OS doesn't have users, but operators. What's the difference? Control. You use a phone, you use a coffee machine, hell you even use a car these days. But you operate a bulldozer, you operate a crane, you operate a plane. You use Linux, you use Windows. You operate Dusk OS.

There’s something I adore about the aesthetic of this project and how this is written. I’ll probably never run the code myself, but I’d watch this show on Netflix in a heartbeat.

It's close to TempleOS in that sense.

After civilizational collapse you won't need computers, you'll need bows, arrows and basic agricultural tools that don't require motors. Forget about electricity or anything requiring it, that will be the least of your problems. Assuming there is still sunlight enough to keep photosynthesis running. If not, you may want to find some place high.

And if you are going to do something like this then you really should make it entirely self hosting and forget about networking and communities and such because by the time you need it those will all be gone and the best you'll be able to do is sneakernet. I hope your USB sticks will be EMP proof... papertape?

Light at night to work was a major achievement for humanity and there are still many lacking it reliably today. It would continue to be useful post collapse. Portable solar panels + lithium battery banks would be quite useful indeed.

They will last a few hundred cycles and then you're done.

Anything that you rely on in a situation like that should be 100% maintenance free and should not rely on technology more complex than what you can make today in your garage with hand tools. Go over that level and it will just buy you a little bit of comfort but it won't last.

Candles: yes. Solar panels+batteries+inverters: forget it. Even a basic charge controller would be a major feat of engineering under those circumstances. I have some ideas on how I'd make that work if I really had to but I'd probably be too busy defending myself from marauders and taking care of my crops to be able to work on it...

What's interesting here is that plenty of the stuff that we take for granted is absolutely state-of-the-art engineering and there is no way that you'd be able to continue to rely on it for more than the available stock that you could still lay your (likely very grubby) hands on for spares, and even those have a limited shelf life.

Bootstrapping to the present day level of technology where working semiconductors can be taken for granted will take a long, long time. Even if you have the knowledge. Assuming that enough of the eco-system survives that you will be able to survive and procreate I'd estimate at least several generations, if we didn't entirely regress to some more primitive state. Mad Max would be an amazing level to bootstrap from, having working computers means you're focusing on the wrong thing. Electric light might be a good idea, but personally I would not count on LEDs or bulbs to last long enough in quantity and if you're going for batteries keep it simple: lead acid flooded cells (You'll need some basic chemistry tools and materials to make sulphuric acid, but that's probably doable).

We can get to working semiconductors in our garages relatively easily. They won't be anything like the integrated circuits we have today.

It would be an interesting hard sci-fi novel to see the village black smith replaced by the village semi-smith and wind mill replaced by the nuclear mill.

> We can get to working semiconductors in our garages relatively easily.

Yes, if you rely on a bunch of hard to obtain tech, a stable energy supply, relatively pure starting chemicals and a whole pile of tooling. I'm aware of Jeri Ellsworth and Sam Zeloof, both are off the charts smart and capable and put the lie to the 'relatively easily' for anything approaching usable devices in quantity.

You'd be better off to learn how to wire relays, assuming you're going to find enough power. It's much easier to get to reliable logic circuitry via that route than it is by homebrew semiconductors, that's an optimization that you can do without if you have nothing. Even tubes would be simpler. And much, much easier to make. Nothing that you couldn't do with early 1900's technology.

Nuclear power in such a situation will likely only kill the operators. Wind power could be relied on for many decades, the Jacobs Wind Electric powerplant that was installed in the arctic still worked decades later under some of the most harsh conditions on the planet. Keeping it simple is key to anything that needs to work in such a setting.

Windmills with simple generators and no (or very simple) batteries could work.

Some electricity some of the time is very useful compared to no electricity, ever.

Or waterwheels, which would work most of the time (unless there's drought) and historically the main non-animal powered motive force for most cultures pre-steam (alright, wind too in the form of windmills and sails, but water always flow in rivers...)

Lithium batteries will be dead within a decade. Solar panels within 3. Think about that the next time you cheer on a huge solar installation as a win for the planet.

Schottky solar cells are pretty simple to couple with thermoelectrics.

Solar panels will work a lot longer than 3[edit: 0] years. I've got a batch here of 16 of them that are from the early 00's and that are still producing ~80% of their rated power. They're not even particularly good cells.

Lithium batteries will indeed likely not last for a decade but that has nothing to do with solar installations, though you could combine them (but you'd have a choice of chemistries and for stationary installations where weight and size are not usually the limiting factors you can do better than Lithium).

Parent comment said solar panels would last 3 decades, not 3 years.

Ah yes. But still: solar panels will just degrade, they will work until they are physically damaged. People replace them because they want more power for a given surface area, usually not because the panels no longer work though there were some pretty bad batches from certain manufacturers where the backing delaminates or where there is water ingestion. Those are no longer safe. But well made panels should last a long, long time. They are warrantied for 30 years, I'd expect them to last a multiple of that unless they get impact damage.

I assume they meant three decades.

Yes, indeed they did, see other comment.

Solar towers have lifespans longer than our own https://clearspanenergy.com/renewable-energy/solar-tower/#:~....

As opposed to fossil fuel infrastructure? I am not seeing your point.

I'm currently using equipment that's working perfectly well running on NiCads and solar panels that are only a little younger than I am. I expect in another 50 years they'll be a bit on the tired side, but it won't be my problem.

It won't be your problem but I'd expect whoever inherits it from you to still be able to get between 20 and 30% or so of the original capacity out of it. Which is likely ample to run simple stuff, assuming the simple stuff itself survives that long.

All the flashlights at my home are powered by ~14 year old 18650 cells scavenged from old lenovo battery packs (I think mostly x61 batteries).

Usually the case is that one cell in a 'dead' battery has far lower capacity, and might only run a flashlight for 20 minutes (which is still useful, though I dispose of those). The other cells are comparable to new ones in capacity (maybe they're 90% or whatever of their original).

We often have a distorted idea of lifetimes and reliability because absent real scarcity we're overly prone to discarding things which can be fixed (e.g. using 5 bad laptop batteries to make four good ones) or are still usable. If you have no other source of light at night a flashlight you can only use for 20 minutes is still a life changing transformative device.

One could imagine searching for battery technology with much longer lifetimes. But I think you'd be better off with the best batteries you can find out of thousands of lithium cells ambient around you than just a few 'super batteries', esp if you're only talking about your lifetime or the lifetime of your children.

I'd bet the US national grid would be back up in some capacity within a year. People tend to assemble in times of crisis.

I'd bet it wouldn't. You might have some tiny local grids but anything city sized would require a lot more coordination than you might think.

Likely there wouldn't be as many services, but don't you think long-range transmission and a good set of scattered plants would provide at least some "modern" facilities within that time? I'm sure the ownership and management structures would look very different as well as usage patterns (probably many services would be centralized in communities).

Think about the infrastructure behind those plants required to keep them running. maybe you could keep a hydro plant running for a while, at least until the bearings fail. But anything relying on high tech, fossil fuels or nuclear would stop long before solar and wind power became unavailable.

Which is probably acceptable to a lot of people, who would have adapted to the new reality by then. Capacity would have to be rebuilt.

The grid is always changing in capacity and output. What makes the grid "the grid" is the ability to transmit power long-range based on demand. I'm saying that that kind of high-level coordination would be a pretty high priority for people in the face of apocalyptic disaster because people tend to be like that. Just trying to counter the cynical doom and gloom that nudges people toward hard-libertarian/ancap/mad-max mindsets.

Have you been in a coma during the past decade? I can’t fathom how you could possibly come to this conclusion after COVID.

Despite all the kerfuffle there were massive mobilizations of people and resources still, and that was even in the face of a silent killer, one that didn't trigger apocalyptic feelings in a critical mass of people as much as, say, widespread EMP attacks from nuclear explosions in orbit that affect every piece of electronics we've come to rely on, or a horrendous chemical spill that changes the atmosphere's chemistry. The latter are much more immediate and appeal to deeper instincts as regards risk assessment than an invisible microbe that only infects a few species of mammal.

The main source of power for early radios in pre-rural electrification 1930s West Texas: small windmills that charged batteries for them - the wind is almost always blowing in West Texas. Zenith Radio Windcharger was the main type I remember seeing at the American Windmill Museum in Lubbock, a must-see if you find yourself in Lubbock for some reason.

On that note, learning CW (Morse Code) well enough to understand it by ear would be one my technical priorities if I were seriously worried about civilizational collapse, as well as more about early (pre-transistor) radio technology.

Yes, windmills would be a good choice.

I've made a whole trip around the USA to make pictures of these old windmills, an amazing number of them are still around, sometimes even in working state. Waterpumpers for the most part but also lots of other uses including electricity generation.

> about early (pre-transistor) radio technology

I dunno if that's a good tradeoff. It's fairly difficult to make good tubes and the performance of pre-tube radio technology is extremely poor. Yet your home alone probably contains hundreds of transistors which could be scavenged to make radio receivers and transmitters.

Constructed correctly these devices have no significant wear out mechanism.

DuskOS is specifically for the interregnum period before full collapse, when microchips are still readily accessible, but can no longer be produced.

(I think the author’s predictions about the collapse of civilization are overblown, but still, if you’re going to engage in a middlebrow dismissal, at least do us the courtesy of reading and understanding what you’re dismissing.)

> (I think the author’s predictions about the collapse of civilization are overblown, but still, if you’re going to engage in a middlebrow dismissal, at least do us the courtesy of reading and understanding what you’re dismissing.)

You should probably do me the courtesy to not make assumptions about what I've read and understood. This is such a common trope on HN that it made it into the guidelines.

The whole idea that the collapse of civilization is something slow is a massive assumption, it could be all over in a few microseconds so fiddling around with tech at a level like this is an exercise in futility: if this is the level that you end up at that's pure luxury and proof that civilization has not collapsed and likely will not collapse in the near future.

Microchips are not going to be much use in any situation where they can no longer be produced. They are the pinnacle of a very large pyramid and without the base the tip is next to useless.

A better - and far more immediate - use case would be to target the re-use of scavenged chips in the present day undeveloped parts of the world which are 100% dependent on imported goods.

Economic war, if it were to become full-blown, could result in locally constrained availability of microchips while civilizations still go on, no? Pandemic x5 or so. What if India and China decide to trade some nukes? It's not likely, but it's a possible scenario.

It's a possibility. But in that case an all out effort to recreate a competition free fab in the west would be able to put something usable together in a few years. I don't see that as a scenario where you'd need to go for scavenged hardware, any old cellphone has enough power to deal with factory automation at that level and well beyond it.

We know exactly what to build and how to tie it all together, the reason we currently do not is economics, not because we could not.

Personally I think the degree to which the Western world is making itself dependent on supply chains like that is irresponsible, but I also think that to some degree it is the same kind of thinking that went on in Germany to make itself dependent on Russian gas. It goes something like 'let's tie ourselves together at the hip in an economic way so that a shooting war is no longer an option'. This presumes rational actors on both sides. Which was a giant mistake in the case of Russia and which may well be a giant mistake in the case of China and Taiwan. But then there is still Korea and possibly Japan.

> But then there is still Korea and possibly Japan.

If china becomes that aggressive that they are in open war with Taiwan, Korea and Japan won't be an option either.

https://news.ycombinator.com/item?id=35046551

"easy to imagine - a large enough collapse that making new ones is very very hard, especially compared to scrounging and repurposing the huge amount of existing ones. Imagine a huge population decline, and search parties looking through the ruins for useful tech, like computers, decades later.

Such a society could probably make new computers, but the cost of making new, much worse specced computers, may not be cost effective for a long time as long as old computers can be found and repaired.

Edit: such a movie or novel would make for great Fallout-ish world-building. Imagine for instance a large hotel in a big city running the booking system on an old laptop in a central secure location, and new 8-bit computers used as terminals at the front desk. Public access 8-bit computer terminals at public libraries connected to copies of WikiPedia running on some old gaming rig and such things. :-) "

Yes, it would make for a nice movie. But real life isn't a movie and of course I'm fine with individuals making their own decisions I still think we should look at the general picture through the lens of finding out what the likelihood of certain scenarios is first. Prepping for an intermediate stage is ultimately not a very good disaster mitigation strategy, you plan for the worst and hope for the best. This intermediary stage of a slow slide isn't going to be stopped because of some FORTH code if and when it happens. But I wish those who believe it will all the best.

edit: I think this sort of thinking is more of an excuse to scratch an itch than a realistic scenario. We are not so addicted to computers - yet - that we could not revert back to some stable society without them, but some of us are so addicted to them that we can't imagine a society without them. The first couple of hours of any major crisis will likely correct that in ways that are going to be very unpleasant. Effectively aiming for something 'in the middle' is aiming for a crisis that isn't a crisis. Anything that actually is a crisis would be better served by acid free prints of Wikipedia and the designs of common farm implements that can be made with just bench tools. You'd be on your feet again in a short few hundred years. But if you remain addicted to playing around in FORTH while the world goes to pieces around you chances are you'll be kicked out of any community you're a part of because you're slacking off your farming duties. Any world where you think you might need this will be so drastically altered from the one you live in today that you won't need it, you'll be way too busy to try to stay alive.

Och c'mon. Agriculture even from 200 years ago can feed much more than 2/3x people than directly working with it. There always will be place for people tinkering and trying new things and or trying to bring old things to life. Computers do have legit use cases like communication, access to knowledge etc... There will be place for comp geeks in post apo world, but certainly not for many. They will have to prove themselves useful though.

Yes it can. But whether or not we'd be able to maintain that level of efficiency is very much an open question, it presumes the availability of 10's of millions of horses and people that know how to use them as well as a ton of tools that we haven't got. Cut the power and you have a problem on a scale that is hard to imagine.

I think some comments here have lost track of how absurdly labor saving simple PLC applications of computers are. There is a lot less farming duties to do when the computers manage irrigation and keep track of schedules.

You're assuming a sudden collapse, rather than a gradual deterioration of society

Yes, because that is (1) more likely and (2) much harder to recover from. A slow and gradual collapse of society can be slowly and gradually reversed and is as long as you can't point at something catastrophic going to stretch the definition of collapse unless it is so slow that it is indistinguishable from change.

'slow collapse' is a bit of a contradiction in terms, all collapses that I'm aware of are sudden, they are events on the timescales that we normally care about, even though if you zoomed in on them enough they would show plenty of detail. Once started they will run their course. Gradual deterioration of society isn't something you need to make contingency plans for. It will happen, or it won't but your FORTH based OS isn't going to make a meaningful difference. For the aftermath of an EMP it will be useless and for a deterioration that takes a year, a decade, a century or even longer there will be other, far more effective mitigation strategies.

But fine, if that's the pretext the OP needed to build their thing I'm perfectly OK with it, I just don't think that's in any way a realistic premise. See also: preppers and shelter fanatics.

The people that are best positioned to deal with all of this are the Mennonites.

My solar powered encrypted radios let me take your basic agricultural tools without worrying too much about your bows and arrows. :D

> hope your USB sticks will be EMP proof.

You don't keep your spare computers in a steel and copper clad safe?

Right next to my supply of tinfoil...

Dr. Strangelove had a different opinion on what was needed after civilization collapse…

Are we currently doing all we can to prevent collapse? It seems like they vaguely try to stop pollution and WW3, but what about the Carrington event? And why does it seem like individuals are far more interesting in preparing for collapse than preventing it?

It really feels like people would rather live in 1800 than continue with industrial society, even if the environmental toll and body count was far higher than what future tech could achieve.

> And why does it seem like individuals are far more interesting in preparing for collapse than preventing it?

That is a great question for which I don't have an answer, but thank you for asking it, it is something to think about.

Based on how many people seem to have a bit of understanding for Kaczynski... I would guess that a lot of people kinda like the idea of collapse....

Or it just doesn't seem like there's much an individual can do.

Prevention requires mass action, preparation can be done by the individual

You ever watch videos where something horrible is happening but all the people in the video who could do something are also recording with their phones?

I think there's a connection here.

I love the ideas behind this. I think too many people are more than happy to jump on the bandwagon du jour without spending a moment to consider how much of the world can't, for some reason or another, participate.

Running NetBSD without languages like Go and Rust clearly shows that compiling and running portable software is perfectly doable. If the Internet should collapse, just having a drive with a full set of NetBSD sources plus all the tarballs downloaded for source compiles in pkgsrc would make it possible to re-bootstrap dozens of targets. That's not possible when we treat the world as a fast amd64 with many gigabytes of RAM.

I'd love to see how far Dusk OS can go.

Neat project. I tend to believe things like this will be more useful 50-100 years post-collapse though. And it seems to me the post-apocalyptic computer niche will be more likely filled by the vast quantities of RPis left in junk drawers rather than the dozen or so computers that will have been prepped with something like this.

Post apo Graphikos - list of addresses of drawers with unused rpis to be used to save humanity by playing doom and block ads in non existing internet, seems like good good artifact for post apo Indiana Jones

Where can I enlist my drawer?

Oh I’m so glad someone has already started working on something like this. This has been one of my biggest concerns for a while

Edit: Are there any easy entry points to this ecosystem for people interested in the concept but not familiar with the details? Like a competition to see who can get this up and running some set program the fastest on microprocessors found in the trash/recycling? Or what’s the most unexpected device people can get this running on?

I too was losing sleep over the tought that we would only have TempleOS after a catastrophic collapse of civilzation.

After TempleOS and SerenityOS and UnityOS, I am going to name my attempt SchismOS.

I am not looking for users, only people willing to commit.

An Arduino would give you a good basic idea of the lowest levels of computing that are still reasonably accessible. Otherwise: PIC chips and repurposed smartphones.

Yeah I was actually thinking about it too.

Strange premise but cool project. I think during a civilizational collapse and after there will be plenty of computers we could run plain old Linux and windows on.

Yeah, in this world, I'd imagine we'd lose electricity before we lose access to computers.

The list of supported devices includes Apple IIe and Sega Master System, which are positively rare artifacts today compared to the glut of smartphones, laptops, smartwatches and TVs. The simple old stuff is hard to come by compared to the integrated modern stuff.

To put it in perspective, a rough estimate is that there were 20 million Sega Master Systems ever sold worldwide, compared to 20 billion (and counting) smartphones.

even if those machines are all you come across, they already have operating systems available for them, with corresponding software that's already been ported. I have a PS2 with a BlackRhino Linux LiveDVD mostly because it's interesting, but I don't anticipate ever having to use it for anything important.

I like the idea of exploring what it means to optimize a system for simplicity.

But the “collapse” backstory seems silly and redundant, and dare I say it reminiscent of the kind of prepper that relishes the prospect of apocalypse, just so that all their prepping pays off.

Seeing civilizational collapse or decline as far fetched is a direct result of your privileged position in the world.

It’s more a result of being old enough to have seen countless histrionic doomsday predictions not come true.

It’s not too far-fetched, it’s happened before: https://en.m.wikipedia.org/wiki/Late_Bronze_Age_collapse

The rationale is silly, but the project itself looks interesting. No need to tie it in to civilization collapse though, an odd marketing angle really.

Yeah, this is an interesting project, but the author's belief that society is going to completely collapse by 2030 is... odd.

If society doesn't collapse by 2030, who's harmed by this project existing?

And if you're going to pick a date, why one over another?

+7 years seems a reasonable goal to make progress towards

By this project? No one, probably. By this "society is doomed, the world is gonna collapse" type of thought? Many people. Loads of people for a myriad of reasons get conned into this dillusion and either (a) stripped out of their hard earned cash or (b) spiral further and further into a dillisuonal state where they either hurt themselves or people around them.

Also it's effin goofy. Nerd CEOs building their underground fantasy post apocalypse shelter, thinking that not only that the society will collapse I their life time into a Fallout-ish World, but that their shelter is going to be of big help. But the goofyness is just that, it's the promotion and prophecies that are harmful.

And at a slightly higher level: the likes of Thiel and company buying up remote property. I wonder if they have figured out how they're going to get there when they have to. In their worldview NZ is probably next door but when the lights go out getting there is going to be a serious challenge, probably much more risky than simply staying put wherever you find yourself.

I'd understand if we were dedicating substantial resources to prepping... but we're not.

The number of people and amount of treasure spent on it is a rounding error.

Let other people have their hobbies.

It's pretty dangerous when those that are in a position to make a difference are also able to position themselves in such a way that they won't go down with the ship if they don't.

That's getting into Illuminati territory.

Thiel is rich and powerful, but he isn't that rich and powerful. There are more than enough counterbalancing forces.

And it starts to be way too thought policy (for me) deciding "You can't be rich and worry about civilization collapsing."

https://www.theguardian.com/news/2018/feb/15/why-silicon-val...

That used so many words to say "Thiel bought some land in New Zealand and thinks the current world is going to end."

And he's removed all the updates where he said the collapse was happening and then turned out to be wrong.

He's basically just another doomsday cultist.

"The reasonable man adapts himself to the world; the unreasonable one persists in trying to adapt the world to himself. Therefore, all progress depends on the unreasonable man."

Pretty much by definition the first people to prepare for any disaster will be people who are or at least appear to be unreasonable.

We benefit from a diversity of opinions, even when some of the opinions are wrong.

This cartoon always comes to mind. People have been saying this since before civilizations probably. Eventually they are right.

https://dustoffthebible.com/Blog-archive/2012/05/25/the-end-...

saying society will collapse by 2030 is more reasonable than saying people will run a niche operating system after it does. Societal collapse looks different to different people but I'd be surprised if the western world hasn't radically declined by 2050.

I mean we already see the issues with international supply chains failing during covid, and we haven’t even really seen the significant impacts from climate change. Yeah, most people on HN aren’t going to really need this in the next 5-10 years as we’re predominantly in locations that will be maintained the longest. What about people in countries that already barely have functioning supply chains? This could be immediately useful for them.

Immediately useful how exactly?

Very low hardware requirements, so basic industrial control at the level where you'd otherwise use an Arduino or so but on scavenged hardware. Forth is ridiculously simple to get an implementation running.

https://github.com/nornagon/jonesforth/blob/master/jonesfort...

Is a nice starting point. It's obviously not as compact as say 'Brainfuck' but it is far more versatile.

"This could be immediately useful for them."

That's the thing: there are already many hackers and makers working on all kinds of ways to reuse old electronic. If they cannot really use this now - who can after a collaps?

Sorry, I’m not following. Are you saying they are able to use this now? Or that it can’t be used now?

The author is saying his project has only use after a collaps. I doubt this project will be useful after a collapse, if no one can make use of it today.

I think the overlap between 'people in civilizational collapse zones without reliable internet access' and 'people with Forth and general computing skills sufficient to expertly operate random hardware' is very narrow indeed. It's a cool project, but the complexity of most hardware is such that solo reverse engineering it for practical purposes is likely a fictional scenario. The chances of an engineering community that can be productive under civilization-ending pressures also seems pretty unlikely.

If you have technology skills and reasonably expect to experience some sort of apocalyptic scenario, ISTM you'd be better off with a solar charger, a few rugged smartphones, and several SD cards containing backups of Wikipedia and as much reference material as you can collect. Getting to a better location (either by traveling or improving a deficient one) is likely to take up most of your mental and physical energy, and electronics are in the 'nice to have' part of your inventory, below weapons, tools, and basic survival supplies.

If you want to go really cheap and low power, consider stocking up on ESP32 devices, which are super-capable and can be had for only a few $ each.

> If you have technology skills and reasonably expect to experience some sort of apocalyptic scenario, ISTM you'd be better off with a solar charger, a few rugged smartphones, and several SD cards containing backups of Wikipedia and as much reference material as you can collect.

That will buy you a decade. I'd go for a paper version of the stuff that you think is important enough to haul around in physical form for many decades and immediately start spreading these around in numbers large enough to ensure that some copies would survive. And even that assumes literacy, some level of schooling to survive, and enough specialization to be able to set aside time for people to study all of this. The best mechanism for something like that would probably be a priesthood of sorts... they tend to be long lived and good at preserving knowledge (and eventually to use it as a power base for domination of the unwashed masses which they work hard to keep dumb but that's already many steps down the line).

Good points. I think it's been linked on HN before, but one of the most robust libraries in the world has its texts engraved on wooden blocks: https://whc.unesco.org/en/list/737/

Deciding what to conserve and making it succinct enough to be manageable but detailed enough to useful actually seems like a great use case for AI. I'm a little more optimistic than you on the timeframe, as I have electronics that are decades old and work fine. Of course, if they suddenly stop it will be a source of disappointment rather than survival anxiety.

Capacitors are the weak point, the electrolyte just goes after some time, faster if they are warm (for instance because they are in working gear). Besides that there is apparently actual wear at the sub-microscopic level from using electronics.

https://en.wikipedia.org/wiki/Electromigration

This puts an upper life-span on VLSI chips and has practical effects on the fabrication processes and materials used. The higher the integration level, the smaller the feature size the more important this becomes.

So if you want something to last a really long time use the largest feature size chips that you can still make your designs with.

I have read about Collapse OS (I think it is because it was about a Forth-based end-of-world OS) a while back, and it has made a lasting impression on me. This might be the reason why I enjoy easy bootstrappable/self-contained toolchain setups and single static binaries. I remember thinking at the time that this idea might be too far-fetched what is more practical approach to this, and remember I also went down the rabbit hole of exploring some Sugar OS or Kiwi something that was used to bring Wikipedia and Stack Overflow to African schools or something similar.

This might sound a little out there, but imagine a world where even if npm/cloudflare/AWS goes down or if there's a world war and intercontinental fiber connections are attacked, everything keeps working. In a strange way, I'm incorporating these ideas into my current project[0]. Thank you weird project, which might have had a much deeper impact on me subconsciously.

[0]: https://github.com/temphia/temphia

https://git.sr.ht/~vdupras/duskos/tree/master/item/README.md

The thought of a solitary hacker somehow saving our electronic heritage with nothing but a Sega Master System brings a grin to my face.

For Duck OS to shine, you really do not need a civilizational collapse. Consider a world that existed in 1960s, where computers were not affordable by a median family and only Universities and Research institutions can have them. We can get into that world with a dash of hyper-inflation and war in wrong places.

Semi-conductor industry has complex supply chains, and if cost of bunker fuel shoots to say $500 dollars a barrel, we are done with transcontinental shipping. There are few ways of getting there -- I know Civilization Collapse has dramatic effect and is the right marketing.

A dash of hyper inflation and some war in the wrong places and you'd be very, very lucky indeed if you ended up in 1960. In that case you'd be in 2023 in five more years. I think the chances of shooting way past that are much higher than being able to stop the slide in 1960. But I like your optimism.

These kind of self-contained start-from-scratch projects are comforting, and can be intellectually interesting, but it would make much more practical sense to me to spec out some simple hardware than can run linux, or a porting toolkit to make it easier to run on phones etc.

The problem is that even if you have a perfectly preserved box of modern cell phones, they're intentionally not really hacker friendly.

Half the time, you have a locked bootloader, and even if you don't, you probably need a bigger machine to stage and load new software. Even when you're done, you're stuck with a form-factor that's of limited alternate utility. Phones make a decent media player or ereader, but they're not really useful for many use cases. I'd expect, once you've got a few devices up and capable of preserving, say, human knowledge as defined by Wikipedia, you start looking for devices that are more GPIOs and less screen-and-storage. Access to something Arduino-esque could be a big force multiplier, because it allows process automation with relatively low manufacturing skill. Even something like an old mechanical clothes washer timer-- open and close a few relays over time, maybe respond to a few sense switches for emergency stop if the tub leaks-- requires a fairly high degree of mechanical tolerances to get right, or you could implement the same functionality in a few dozen lines of code for an Arduino-esque device, as long as you have the GPIOs. Phones are hardly amenable to that-- you try soldering a big relay to a tiny BGA processor pin.

Some of the documentation seems also focused around the idea of salvage at the component level. This may be a more feasible option too-- when you start to look beyond "computer/phone devices", there are a lot of embedded Z80s and EEPROMs that you could desolder and press into service, and they represent a lot more of a consistent baseline than 400 chips that are all theoretically ARM family, but have wildly different memory maps, peripherals, and varying instruction sets.

Right it should be mandatory for phone makers to release their linux changes/drivers etc when they retire devices from sale.

https://web.archive.org/web/20230716210954/https://duskos.or...

Glad it uses “almost C” instead of “almost C++”. There is zero chance I could debug compiler errors for modern C++ without the web. With C and K&R, I have a fighting chance.

If civilization collapses and you live on your own, you will not have much data to store, organize, etc. Therefore a computer won't be that useful to you.

What will be a priority to you will be to survive, which essentially means how to get drinking water and nutrition, as well as territorial control and self defense.

A computer will be useful to you if you become a ruler that is responsible for the census, taxation, etc.

Having a large collection of offline reference books, maps, and music/movies for entertainment would still be desirable for those who can find a way to power/charge a device.

Rather than building a 'collapse OS', maybe just hoard some reference material, and have a plan (and multiple backup plans) for powering/charging devices in the absense of a power grid or fuel for generators/vehicles.

I don't think anyone's going to be doing much coding in a post-collapse world, let alone tinkering with the internals of an OS

Maps will be useful but after EMPs and nuclear winter, consuming maps on a powered device will be impractical.

Paper maps will be far more practical.

First targets going down in a global war will be GPS, Galileo, Glonass, etc. so all those systems will go offline right away.

All power will go to water pumps, heating, air conditioning, mining, etc.

re: maps — I accumulated a lot if regional and national maps ( thank you AAA!!) but they do go out of date.

Major challenge with prepping is uninterrupted vigilance maintaining your supplies… which likely is an Amazon marketplace business in and of itself… lol.

There are apocalypse shelter services for rich people that take care of this for you.

Wayback Machine snapshot for anyone else getting the "exceeded bandwidth" error: https://web.archive.org/web/20230716210954/https://duskos.or...

> "The first one is when the global supply chain is broken and the second one is when, in a particular community, the last modern computer dies."

Can someone please explain to me how modern computers will die? I don't see how solid state electronics will break, will they eventually rust in 100 years or something?

Not my area, but at least "big" electronics will probably die because of whiskers that grow until you've got raandom shorts in circuits.

Not sure if integrated circuitry suffers from this or other similar phenomenon to.

https://en.wikipedia.org/wiki/Whisker_(metallurgy)

I've been thinking a out civilization collapse for a while, and to me the greatest think anyone should be building is simply a computer with wikipedia installed, running on solar batteries (or something even more durable). Meant to be working for a 100 years

If you’d rather development to last 100 years, may I suggest GNU Hurd?

I'd suggest printing it out on acid free paper and guarding it with your life.

You've probably already seen it and it doesn't fit most of the requirements but some might be interested to see: https://en.wikipedia.org/wiki/WikiReader

I for one look forward to the post computational era … no monthly cloud storage fees, no data backup anxiety, no cookie tracking pop-ups to click ‘yes’ to … back to when I had to remember where I was driving and when the destination would close …

At least there will be no throttling due to cloud costs after the collapse.

So an OS that’s lets user operate very close to metal/driver level? I still can’t imagine what the os will look like vs regular Linux

It doesn't have a GUI so it doesn't 'look' like anything other than a way to define new FORTH words and some basic low level drivers.

It does have a GUI. Not much is written for it yet, though

=> https://git.sr.ht/~vdupras/duskos/tree/master/item/fs/doc/sy...

That's not a GUI. That's barely a display driver.

TIL not only that this is an idea that's out there, but it's a genre (multiple flavors). Neat.

Based on the unavailability of the website now, I'll hold my reservations about their claims.

They're expecting web server loads to be much lighter during the first stage of civilizational collapse.

Is a Minecraft port possible? That could measure the post civilization applicability.

i don't see the point of this, seeing current operating systems will still exist and there will be more of a desire to get those running and supply the hardware for them, than this niche OS.

what advantages would this have over something like IDK, debian?

This OS runs on garbage electronic, you would not suspect, that they have a CPU in the first place.

Debian needs a little bit more.

It's much, much, much simpler and more compact.

The core system is about 8000 lines of code. The C compiler is 1400. Its memory footprint is extremely small (1MB to have a C compiler running on bare metal in the PC architecture). From here, you have a system capable of building C programs as well as bootstrapping itself.

This is level of simplicity is staggering and empowering -- it is completely reasonable to read and understand every single line of code in Dusk OS. This would be incomprehensible in something like Debian.

How are these advantages in a post-apocalyptic world over any other OS that most people could actually use and do useful things with?

I think you are missing the point. You might not have access to any modern hardware + modern, multi MLoC operating system medium/ftp server/whatever to boot in a post-apocalyptic world. This would at least provide some level of computation on simple devices.

Even if you do have access to a modern computer which is not fried, you still need a whole bunch of devices, mass storage stuff and mains electricity to be able to boot it, and then on to find and boot some software.

The OS above would run in things like graphing calculator cpus, 8-bit video games and other "almost non-computer" devices which actually have simple but useable CPUs. It would be better than going back to the 1900s.

I would frame it differently, "collapse" can mean any number of conditions that exist in the world today -- internet access that is limited, censored/liable of being shut down, or nonexistent. Same with electricity. We also rely on institutions that may or may not represent your interests to build software that you rely upon (e.g. the decline of search, the locking-down of major social media platforms, etc), and so on.

Virgil's "collapse" mindset, to me, is about building resilience and independence, and IMO Dusk OS is fantastic at this purpose.

If you're writing it in Forth, why bother having a C compiler?

There is a lot more software that currently exists written in C than in Dusk OS's forth dialect. The C compiler allows Dusk to run this software (once it's ported over)

The FORTH core allows Dusk OS to be ported with an absolute minimum of work. It's like an abstract machine language.

very good timing of the post, second season of Foundation just started streaming :D

Apple won't allow it.

should really use apl instead of forth.

Much larger hardware requirements.

And brain requirements. Forth is frustrating to me, but APL is just all too much.

APL is super powerful, I got to the level where I could 'read' (decode) it but I've never managed to get proficient at a level where I could write a significant chunk of code in it. Say a line or two ;)

FORTH is extremely easy to get into, a very powerful metaphor that has a ton of parallels in other languages, even if you don't realize it. There are strong connections between FORTH, LISP and shell programming languages to the point that they are in a manner of speaking different ways to express the same thing. Of those three I think LISP is the most complete and feature rich, and FORTH probably the least because it is much more barebones. But that upside of that is that it also requires absolutely no supporting infrastructure, the most basic access to a machine at the hardware level is enough to bootstrap to a usable system in an absolute minimum memory profile.

This is like TempleOS but for collapsists

TempleOS is at the rocket science level compared to this.

But a 32 bit FORTH implementation may well have embedded uses in the present (though, in some circles building something like this is more like a rite of passage and anybody able to use it properly is likely able to re-implement it, possibly better).

Reply on news.ycombinator.com