1. 50

  2. 15

    Google has a project to create a “backpack” of software that should be enough to restart Google in case everything shuts down. Since everything hasn’t been powered up from cold-boot before, we don’t know if it’s even possible. This is especially hard with dependency loops.

    1. 10

      I look forward to receiving ads from backpack google and being tracked after the apocalypse. I’m sure nothing bad can come from that.

    2. 13

      It’s cute and I have to respect the effort put into it, but be honest - I doubt this will be useful in a post-collapse scenario.

      The kind of people who have Z80s and know how to work with them are likely the first ones gone in such a (morbid) scenario, and their usefulness is limited. I don’t think getting semiconductor fabrication would be a high priority in such a scenario.

      If computers are useful post-collapse (big if), people will scavenge x86 boxes because they’re everywhere, people are far likelier to know how to use it, and they already run most everything. There’s enough around that scavenging is easier than using 1970s tech, and the older examples are quite workable with a soldering iron anyway.

      1. 8

        Maybe the author is completely earnest, but I prefer to think this is a sort of RPG, broadly in the “zachlike” genre.

        1. 6

          The assumption that all computers will break down in two decades is also quite wrong. There are still plenty of x86 boxes built two decades ago that are still working fine, or suffer from trivially repairable problems (like cap plague). Lots of boxes from three decades ago are still working as well.

          That’s not mentioning 90’s workstations that were absurdly expensive at the time, but were really built to last. Lots of Sun/SGI/HP/DEC boxes are still lurking in labs, manufacturing facilities, and other places.

          I suppose a more relevant post-collapse project would be an easy to make dumb terminal system to enable sharing the remaining machines.

          1. 3

            I suppose a more relevant post-collapse project would be an easy to make dumb terminal system to enable sharing the remaining machines.

            Probably more radios (tube, not transistor).

            1. 1

              The process nodes of the old stuff last longer. The newest stuff breaks really fast. So, it’s also possible we run out of new stuff before the old stuff. Optimizing for the old stuff then makes even more sense.

              And making backup stuff on old nodes, too.

              1. 2

                I don’t think this assumption is universally applicable. There are more cheaply made devices now because it’s finally cheap to produce them. Whether the good ones made today are going to last, time will tell.

                There are reliability bitter spots, like the late 90’s generation of chips suspectible to electromigration, before anyone figured how to deal with that issue. Early to mid 00’s boards before solid-state caps were especially suspectible to cap plague, from my “anecdata”. In a hypothetical societal collapse scenario, we’ll probably end up with a few narrow bands of device generations rather than a contiguous spectrum, but then, if you account for legacy hardware outside of the IT industry, it’s probably already like that.

            2. 1

              Known to run on:

              • A RC2014 through a serial link. It can also have a PS/2 keyboard directly plugged in.

              • A Sega Master System or a MegaDrive (Genesis) with video output and D-Pad input and/or a PS/2 keyboard adapter.

              I have absolutely no way to run this now, as I sit here in a room full of capable computers and there being no immediate (obvious, as there would be in a ‘collapse’) threat to my ability to live. I think your assertion that x86 is that way to go, is the way to go. ARM is probably more common at this point, but the architecture is a mess in terms of backward compatibility.

              1. 1

                ARM is probably more common at this point, but the architecture is a mess in terms of backward compatibility.

                Yeah; most ARM chips are in embedded or otherwise very locked-down systems. If you can find a hard drive with your (if you can find it) working example of a system, you’re good to go and can reconfigure it as needed.

            3. 10

              Post-collapse FORTH will rise like Pheonix from the ashes, and the Neo-Chuckians of Stack Valley will be the builders of a new, and wonderous age.

              (Very cool OS here, I will have to look through the source. I’ve been meaning to learn Z80 assembler by example.)

              1. 6

                It’s been a while since the 70s. I wonder if there have been any chemistry advances–within hobbyist reach–for making a larger, slower, less energy-efficient processor using a less precise and less hygienic process?

                1. 2

                  Man, that reminds me of an idle thought I once had: I wonder how many hobbyist genetic engineers there are out there just mucking around with whatever is available to them. Apparently it’s a thing:


                  So to keep the thread on topic are we going to get a Fallout future or a Bioshock future?

                2. 2

                  Kind of ironic the site’s down, lol

                  1. 2

                    Isn’t that what NetBSD[0] is for? I mean it doesn’t quite go back to working on a Z80, but it does go back to the i486, and the 68k CPU’s (and a whole bunch of other CPU architectures[1]). So the chances of finding a machine NetBSD can run on is pretty high I would imagine.

                    0: https://netbsd.org/

                    1: https://netbsd.org/ports/#ports-by-cpu

                    1. 2

                      Both z80s and 6502s are still in production.

                      Dropping down a level, custom cpus assembled from ttl is a thing and could be used as a bootstrap starting point (many on hackaday, but see https://eater.net/8bit for an example).

                      As for practicality, think how much your survival rate goes up when you can track supplies, trends and other things even an 8bit machine can do not to forget calculating ballistics for the local warlord ;)

                      1. 1

                        Clearly from most of these comments someone should design an open source z80 clone with an instruction manual on easy fabrication with limited resources

                        1. 1

                          I’d buy that kit. I’ve been skimming the retro and SBC designs for something with a board and BOM under or about $100 that’ll run CP/M and other software from the 1980’s.

                          1. 2

                            There are a bunch (eg: https://www.tindie.com/products/tindiescx/sc126-z180-sbc-motherboard-kit/) Small part count z80 systems (4-6 chips) are possible on a bread board (See hackaday).

                            1. 2

                              Good find. Bought one :)

                        2. 1

                          Is z80 a reasonable architecture for such a project? While maybe on the power side it makes sense, these days (fully compatible) z80 seem to only be relevant in the collectors scene and various environments (embedded systems) moved to or are moving to other architectures, which seems to make it hard to come by after a collapse.

                          But maybe I am missing something.

                          1. 1

                            There’s a motivation on the site. One reason is that for these old 8-bit chips the transistor count is very low, so it would be easier to recreate them with lower-tech machinery.

                            However, in that space I would think 6502s would be more common to scavenge, as those are in so many machines of the 8-bit era.