Collapse OS (which we talked about 4 years ago) has a successor.
Dusk OS is a 32-bit Forth and big brother to Collapse OS. Its primary purpose is to be maximally useful during the first stage of civilizational collapse, that is, when we can’t produce modern computers anymore but that there’s still many modern computers still around.
Unable to access Dusk OS page, they’ve reached their daily limit. 🙁
I get the same result, according to the wayback machine duskos.org redirects to the following link…
https://git.sr.ht/~vdupras/duskos
Oh, the irony…
Knowing people they’ll simply use Windows XP or whatever old version that still works on their hardware. This project is more of a digital version of apocalypse prepper’s daydreams than a realistic solution. A classic solution looking for a problem if you will. In reality you’ll still come across millions of x86 bits of old hardware, and Windows software. If anything trying to run this will just make life much harder for the user.
Yes, this is pretty much just mental masturbation. I guess if you were really looking at how to maintain computing after a complete collapse of the internet you’d want some sort of prep kit that included a minimal set of compact flash and cfast cards that could boot on as many different things as possible and reproduce themselves to other storage media without introducing errors. It’d be kind of interesting to try to produce flash media that would boot on several different architectures at once.
You’d presumably also want to give some thought as _why_ you’d want a computer in those circumstances, I guess a big technical library on agriculture, medicine, chemistry and manufacturing might be a good reason.
Jeeves,
It would be so awesome if we had standards that enabled this, but it’s a pipe dream. So many devices today make it very difficult to boot up your own code even without having to support several different architectures at once.
The CollapseOS paper is extremely interesting and targets even simpler devices like micro-controllers. These are fairly well standardized within respective product brands, though I think some of the complexity is hidden behind some compiler magic. As long as we can run the compiler somehow, it means we could still use those chips. But I’m not sure how accessible those compilers will be given the events that are envisioned. The point of the project is that modern computers are gone.
Part of the project is dealing with a world where we do not have working computers and monitors. In such an event, it would probably be worthwhile to invest in printed libraries, an archaic concept to be sure, but there’s no question it would be viable without having to rely on electronics. Without an electric grid, even working computers are going to be hard to run. Solar panels could help with that, but paper or microfiche could be better for a hypothetical world where computers have collapsed.
Not to mention 32bit is kind of stupid as we’ve had 64 bit chips for how many years now? Heck even the ancient Phenom II X4 box I use as a media tank and my AMD netbook from over a decade ago sport 8GB of RAM. I seriously doubt if we were to have a collapse you are going to find many 1990s PCs still kicking it, more likely it’ll be laptops from the 2010s where you could replace the batt and all of those were 64bit.
Debian would be a better alternative, runs on almost anything, is very stable, and it’s easy to setup your own local repositories and modify yourself.
NetBSD would also be a good choice. I don’t know. Minux, would rate highly for what he’s trying to do here. I think a simpler design might be better for many use cases, but I’m not sure forth is the best way to get there. C certainly isn’t either, but I understand the appeal of a language which has so many compilers just out there. You’d want something that is understandable, and extendable while also have an easy learning curve and low energy consumption as the grid might not exist as it currently does. A raspberry pi based netbook might be an excellent hardware base, with software unknown.
Bill Shooter of Bul,
The ARM environment is a bit notorious to program for in general. But as far as they go I agree a raspberry pi has lots of merit. It is functionally dense and an energy efficient hardware platform. However even during good times RPI has suffered systemic out of stock conditions for years. I’ve been on the market for a RPI 4 and have yet to see it in stock. This makes RPI unsuitable for this project. If things were to collapse, the chances of finding an RPI layout around are virtually zero.
The best choice, in terms of both abundance and programming accessibility might actually be x86 laptops. Other hardware might have technical merit, but it’s hard to beat the x86 on both ubiquity and ease of programming. Hardware that is more efficient and abundant exists (ie old android phones), but unfortunately it won’t be useful when it’s so proprietary and locked down. In this scenario, we need to be able to repurpose hardware using minimal resources. Even for me today I find it a shame our mobile devices are so restricted that it limits their utility in DIY IOT applications.
Isn’t it just sad how Google has made phones disposable? They could easily force standards upon the industry so that drivers from one version could easily be carried over to another (or a different OS) but Google only cares about eyeballs so the mountains of E-Waste grow ever higher.
Its just a damn shame that my AMD netbook from 2011 is happily humming along with a fully updated Q4OS and can stream music and 720P video while next week I’ll be taking yet another octocore Android phone and throwing it in the trash because without OS updates and without a way to change the OS its simply garbage. I used to try to find uses for them but how many MP3 players can one person need? I already have a nice octocore no longer supported phone for that, can’t find any other uses for the things so into the dump it goes.
That is why I broke no bad talk about X86, because say what you want about how much less power ARM uses but I can take any old Windows laptop or desktop no matter how old and slap a Linux on it and the thing will happily purr with a secure OS that is fully patched, but the SECOND the OEM quits supporting an ARM device? Straight to the dumpster.
bassbeast,
I consider us fortunate that x86 evolved as it did, but those circumstances may have been rather unique. I’m afraid that the path going forward is the ARM model with vendors doing their own thing and not working together for interoperability sake. Why should they fix it? They don’t even want customers installing non-OEM operating systems anyway.
The OEMs have been (rightfully) criticized over their lack of long term support. But I honestly think expecting long term support from them is a bit misguided. Their whole business model revolves around selling new hardware. I think the most ideal solution is really what x86 had all along: strong platform standards that make it possible for users to install whatever the hell they want without being tethered to the OEM for updates. Users should be able to install new supported operating systems themselves on 10+ year old hardware, just as you did with your laptop. There is no technical reason this ability needs to be unique to x86 laptops. The problem is that vendors like the new status quo. Throwing devices away fits in perfectly with planned obsolescence 🙁