Slashdot is powered by your submissions, so send in your scoop

 



Forgot your password?
typodupeerror
×
Open Source Data Storage Earth Programming Software Technology

GitHub Places Open-Source Code In Arctic Cave For Safekeeping (bloomberg.com) 50

pacopico writes: GitHub's CEO Nat Friedman traveled to Svalbard in October to stash Linux, Android, and 6,000 other open-source projects in a permafrost-filled, abandoned coal mine. It's part of a project to safeguard the world's software from existential threats and also just to archive the code for posterity. As Friedman says, "If you told someone 20 years ago that in 2020, all of human civilization will depend on and run on open-source code written for free by volunteers in countries all around the world who don't know each other, and it'll just be downloaded and put into almost every product, I think people would say, 'That's crazy, that's never going to happen. Software is written by big, professional companies.' It's sort of a magical moment. Having a historical record of this will, I think, be valuable to future generations." GitHub plans to open several more vaults in other places around the world and to store any code that people want included.
This discussion has been archived. No new comments can be posted.

GitHub Places Open-Source Code In Arctic Cave For Safekeeping

Comments Filter:
  • by genessy ( 587377 )
    Well, that's an unusual disaster recovery plan.
    • Re:DR? (Score:4, Insightful)

      by presidenteloco ( 659168 ) on Wednesday November 13, 2019 @07:06PM (#59411838)
      Should not rely on a single place on Earth. That's dumb.
      Coal mine? Flood? Really big asteroid? Perma-frost melt?

      You need a globally distributed, highly redundant (many copies), self-migrating data-store system.

      Self-migrating means it periodically checks if there are enough copies of itself, and if not, seeks new storage nodes to copy to, on newer hardware, and reliable-network-presence-verified hardware.

      Such a "Perpetually Persistent" data storage network (open source, not privately run, of course) might stand a decent chance of long term survival of content.

      (Super-ironically, I was working on a prototype of that in Java 20 years ago or so but lost multiple copies of the source code to 3 simultaneous hard drive failures after a turning computers back on after a holiday. )
      • From decades of experience, I'll point out that getting it out of the main data centers into some off-site facility protected with an air gap is the first, critical step. It reduces the risk of someone accidentally deleting the metadata, online, or of wiping the underlying storage system's filesystems. And the system you describe is also why you put the multiple copies on slightly different physical media. Many of us remember "IBM Deskstar" failures wiping out entire RAID arrays at a time.

        • by mysidia ( 191772 )

          also why you put the multiple copies on slightly different physical media.

          It sounds like the people in the article have THAT problem solved by putting the media into sealed containers that can be read with a magnifying glass.

          Only question I would then have is.... How are you going to FIND a piece of work From all the hundreds of thousands of containers full of these plastic boxes... How will you locate a specific software program that you need to retrieve, Or if you pick a box at random... how wil

          • by Kjella ( 173770 )

            Personally my thought is that either you have a high tech society that can produce the hardware and this software is little more than a historical curiosity or you're a low tech society that can't really make use of software that assumes gigahertz processors with advanced instruction sets and gigabytes of memory anyway. Charles Babbage made a lot of the theoretical work for a computer in the 1820s. Theoretical transistors were designed in the 1920s, but semiconductor materials of sufficient purity weren't a

            • There is enormous space in between. A few fools, especially aggravated employees, flushing or corrupting the active data is a very real risk for any large scale software repository.

          • Finding the data is a problem. The concern I expressed was for unexpected catastrophic failure of all of one class of medium at one time, typically due to a commonplace manufacturing flaw. I've encountered too many times where an entire class of hardware had a common flaw, sometimes one addressed without revising the model numbers in the next release of the hardware. Going back and finding the bad hardware for replacement became quite an issue.

      • You need a globally distributed, highly redundant (many copies), self-migrating data-store system.

        Torrents would be simpler. As long as at least one person has a copy and an internet connection it's fine.

      • by tsa ( 15680 )

        Make a DNA strand out of it, put that in a few common ant species and there is the best backup system you can imagine.

      • by mysidia ( 191772 )

        Should not rely on a single place on Earth. That's dumb.

        Just because one site isn't the perfect scenario does not mean they shouldn't start building it.
        They can start working on Site #2 buried on the moon, and Site #3 somewhere else in the solar system later,
        But the costs go up exponentially, and what can they do to even make sure humans would be able to find the site later;
        let-alone have means to regularly dump new archives for preservation?

        There are already other places this data is at -- this

    • Re: DR? (Score:5, Interesting)

      by BarbaraHudson ( 3785311 ) <barbara.jane.hud ... minus physicist> on Wednesday November 13, 2019 @08:44PM (#59412086) Journal
      The death of the Internet, and the loss of all the open source software in the world, is nowhere near the biggest threats to humanity.

      No new antibiotics will put us back in the time when any surgery was a major risk of infection and death. Life expectancy was 50 at the beginning of the 20th century.

      Global ecosystem destruction, along with all the horsemen of the apocalypse, is also a bigger threat than the loss of the internet and open source software.

      A nearby supernova (100 light years) will sterilize the planet. There's planet-killer asteroids. Death of life in the oceans, where 80% of our oxygen comes from (the Amazon is a much smaller contributor). Nuclear war. Biological warfare. Grey goo from nanites designed to do atom-by-atom recycling of garbage that get into the wild.

      The loss of the internet is nowhere near as bad, especially since other networks (local, wifi mesh) will continue to work.

  • Isn't that where Eclipse, OpenJDK and other significant software is from?

  • It doesn't matter how many sites they operate. If they're owned by Microsoft, they're tainted and will ultimately be used to the advantage of Microsoft and the detriment of the community.
  • Re: (Score:2, Funny)

    Comment removed based on user account deletion
    • by mysidia ( 191772 )

      because somehow icy mineshafts are technically superior to a distributed, fault tolerant, revision control system that places code in the hands of hundreds to thousands of people all at once.

      Hey, do you have even a cursory understanding of what
      apocalypse AND safeguard the world's software from existential threats means?

      Those copies in the hands of hundreds to thousands of people all evaporate almost instantly when the nuke goes off obliterating most of them directly and the EMP takes about 5 s

      • by MrNaz ( 730548 )

        So, some disaster strikes obliterating every digital storage device containing along with most of the people who use it, and you think having software backed up on microfilm in the fucking arctic will be even remotely useful? How will you get there? How will you load the software onto a computer? Where will you even get a computer? What use will it be given that the internet will likely be non-existent?

        Something something cursory understanding dumbfuck.

        • by mysidia ( 191772 )

          How will you get there? How will you load the software onto a computer?

          Most likely its Not for you nor the immediate survivors of the disaster. Its for the descendants of their descendants' descendants' great great great great great great great great great great great great great great great great great great grandparents.

          Why else would they be targeting media designed to last 750 years and finding conditions for it where its likely to last for more than 10,000 years ?

          This will be stuff for academics

      • I get what you're saying, but the thing that strikes me here is that if all the technology to use that code is destroyed as badly as you describe, then it'll be at least a couple of generations before it's ever recreated again and the originals read and understood well enough to use. The chances of having the exact same computer architectures, CPUs and other things seems very slim.

        That's not to say that if someone dug up the dinosaurs T-nix kernel code right now we couldn't have some fun looking at it, but

        • by mysidia ( 191772 )

          The chances of having the exact same computer architectures, CPUs and other things seems very slim.

          That IS okay. Recall, that most of the basic tools are written in higher-level languages such as C, and the GCC compiler itself is written in C.

          Applications can and have been compiled and executed on brand new completely unrelated architectures; there's some initial work to bootstrap the compiler for a new platform.

          The huge volume of software that's already written is useful enough that someone would b

    • by leonbev ( 111395 )

      That said, I don't think worst case scenario that it would really work. If the apocalypse happens and society crumbles, the survivors are just going to raid these caves for supplies and probably burn the backup tapes for warmth.

    • {...} do you have even a cursory understanding of how Git works? {...} a distributed, fault tolerant, revision control system that places code in the hands of hundreds to thousands of people all at once.

      Or, put simply from the mouth of the guy who eventually invented Git a decade later:

      (Microsoft just proved to be wimps. :-D )

  • by the_skywise ( 189793 ) on Wednesday November 13, 2019 @07:23PM (#59411868)

    Nat Friedman buried the reel to reel tape containing the copy of the repository beneath some dirt and set a marker upon it - yup, this will be perfectly safe and preserved here for millenia!

  • But also not the dumbest time capsule. It should prove entertaining to someone decades from now.

  • Permafrost? (Score:4, Informative)

    by hawguy ( 1600213 ) on Wednesday November 13, 2019 @08:05PM (#59411960)

    ...in a permafrost-filled, abandoned coal mine.

    Permafrost is not so "perma" any more:

    https://www.theguardian.com/en... [theguardian.com]

    • ...in a permafrost-filled, abandoned coal mine.

      Permafrost is not so "perma" any more

      Permafrost was mentioned and global warming/climate change wasn't mentioned until the SEVENTH thread!

      There may be hope for Slashdot as a place to discuss technical issues, without constant Cato-the-Censor style harping about P.C. issues, after all. (Assuming the delay wasn't because the SJWs were all busy watching the first day of the impeachment show.)

      But hawguy does have a point: If things warm enough, permafrost isn't a

      • by hawguy ( 1600213 )

        ...in a permafrost-filled, abandoned coal mine.

        Permafrost is not so "perma" any more

        Permafrost was mentioned and global warming/climate change wasn't mentioned until the SEVENTH thread!

        There may be hope for Slashdot as a place to discuss technical issues, without constant Cato-the-Censor style harping about P.C. issues, after all. (Assuming the delay wasn't because the SJWs were all busy watching the first day of the impeachment show.)

        But hawguy does have a point: If things warm enough, permafrost isn't all that perma.

        So, in addition to the open source time capsules losing the below-the-permafrost part of their protection, all those quick-frozen mammoth corpses will rot, so we need to get samples before that.

        The only person that mentioned Global Warming here is you, I just pointed out a case where what was thought to be safe permafrost, wasn't.

        You can draw your own conclusions about whether or not you think it's attributable to global warming, a single incident, even an unusual one doesn't really prove anything either way.

        Here's the facts as reported by the seed bank, note that they didn't blame it on global warming either:

        But soaring temperatures in the Arctic at the end of the world’s hottest ever recorded year led to melting and heavy rain, when light snow should have been falling. “It was not in our plans to think that the permafrost would not be there and that it would experience extreme weather like that,” said Hege Njaa Aschim, from the Norwegian government, which owns the vault.

        • It was not in our plans to think that the permafrost would not be there and that it would experience extreme weather like that

          Which is frankly, criminally stupid on their part.

          • by hawguy ( 1600213 )

            It was not in our plans to think that the permafrost would not be there and that it would experience extreme weather like that

            Which is frankly, criminally stupid on their part.

            Criminally stupid? You want someone to go to jail for underestimating the rate of Global Warming - entire governments (well, one government that I know of) continue to claim global warming is a conspiracy theory, so it seems like a stretch to call for criminal prosecution for someone that didn't account for it 20 years ago.

            • This reply makes me wonder if we're up against a language barrier.
              Criminally stupid is a figure of speech (in [American?] English) meaning extremely stupid.
              Second though, that the Arctic would be the fastest to heat up was well known even 20 years ago.
              Hell, it was highly suspected almost 50 years ago. So yes- criminally stupid.
        • The only person that mentioned Global Warming here is you, I just pointed out a case where what was thought to be safe permafrost, wasn't.

          Better and better! Thanks for not mentioning it. ;-)

  • Hardware? (Score:4, Insightful)

    by kbahey ( 102895 ) on Wednesday November 13, 2019 @08:25PM (#59412022) Homepage

    This is a good move, but incomplete ...

    What is needed is hardware that can read the media (e.g. DVD reader, or USB port, ...etc.), as well as hardware that can execute the code (that means a CPU, with RAM, and some sort of a disk, as well as a screen and keyboard at a minimum).

    Why? Because history said so.

    Example: 9 track tape used to be the common way to archive programs and data on mainframes, up until the late 80s or early 90s.
    How many working 9 track tape units are in existence now?
    How many computers can directly compile and execute the programs from the tape as is?
    We are not even talking 30 years ...

    Same thing, whether it is a DVD, hard disk backup or any other media.

    • It's a make work publicly stunt to justify someone's salary. In other words, doesn't matter that it won't work.
    • by oaox ( 1756044 )
      Include DIY reader instructions https://youtu.be/hz170R51w7k [youtu.be]
    • by Shotgun ( 30919 )

      Dude, I can't understand what the engineer I work with did to add a function to the program that I originally wrote without a code review. Screw the hardware. Even with the source code, you're going to have people that don't understand that particular programming language trying to understand the data constructs and methodologies end logic constructs of a lot of uncommented code.

      If it happened after an apocalypse, the people trying to understand this code are going to be the same people that are trying to

      • by kbahey ( 102895 )

        the people trying to understand this code are going to be the same people that are trying to survive. I doubt they'll have the time to build a processor fab to generate an x86 processor, so they can write a compiler, IDE and debugger

        You nailed it.

        I always said that up to the Victorian era, or even the first few decades of the 20th century, there was nothing that cannot be manufactured in a good workshop. Skills and some tools to build further tools, and all is achievable. That includes internal combustion e

  • by PPH ( 736903 ) on Wednesday November 13, 2019 @08:40PM (#59412068)

    ... a code freeze?

  • I guess ICE finally found a way to work with Github better, they provide the children and cages, github provides the servers.

  • by Anonymous Coward

    I mean, come on. That’s the obvious place because no one goes there, because it’s all cold and there’s nothing there. In like, 50 years, it’ll be practically tropical, people will flock there, and when some future people stumble upon that, they’ll say to themselves, “dafuq izdis?”

    Another likely remark that could be made by someone finding it is, “kanya yoosit fura wipin?” and someone else might try to bite it, and conclude, “snot food.” and

  • print it out (Score:4, Interesting)

    by Chicken_Kickers ( 1062164 ) on Wednesday November 13, 2019 @09:53PM (#59412268)
    They should print out the code and store it tomes of books. We have good experience in preserving books compared to digital media. We have books that are 500 years old.
    • by WallyL ( 4154209 )

      Then in 5,000 years we can have people arguing over the canonicity of the Linux kernel! "In kernel/sound/usb/snd-usb-audio.ko#53, the first character is actually a bang instead of a octothorpe, which changes the whole meaning of the code block." And who gets to decide which drivers are canon and which are apocryphal?

  • Start Over (Score:4, Interesting)

    by ShoulderOfOrion ( 646118 ) on Wednesday November 13, 2019 @11:29PM (#59412434)

    Most of these hardware and software standards were kludges built on top of hacks. My advice to future generations after the great Apocalpyse: refactor the whole mess.

  • That's a nice way to say, "...threw the Linux and Android source code down an abandoned mineshaft."

  • by sad_ ( 7868 )

    and it's already outdated the second it was copied from github.
    there must be a better, automated way to achieve this.

  • 5,000 years from now the bird people are going to find systemd and declare the monkey civilization less advanced than previously suspected

  • no, civilization depends on telco gear running proprietary wares and certain networked mainframes for banking, trading and insurance also not running open source.

    c'mon, we don't trust our MONEY to those fucking hippies...

  • 20 years ago open source was already beginning its rise.

A committee takes root and grows, it flowers, wilts and dies, scattering the seed from which other committees will bloom. -- Parkinson

Working...