• Simon Müller@sopuli.xyz
    link
    fedilink
    English
    arrow-up
    79
    arrow-down
    11
    ·
    7 months ago

    I mean, to be completely fair, that’s how data storage works.

    We cannot really just make data disappear, so we let it get overwritten instead

    • mojo_raisin@lemmy.world
      link
      fedilink
      English
      arrow-up
      44
      arrow-down
      20
      ·
      7 months ago

      But clearly the data is not overwritten and this was intentional. How do I know? Because that would amount to a massive amount of data, if it was de to a bug in Apple software or underlying filesystems, it would be detected in monitoring systems “Hey, we’re using 10x the data we should be, maybe we should look into it”.

      The mistake was in the flag code that was supposed to fool us.

      • Simon Müller@sopuli.xyz
        link
        fedilink
        English
        arrow-up
        52
        arrow-down
        3
        ·
        7 months ago

        no when I say “overwritten” I mean that the area is set as deleted in the filesystem and the next time something writes to that area the data that was there before is disregarded.

        • barsoap@lemm.ee
          link
          fedilink
          English
          arrow-up
          1
          arrow-down
          1
          ·
          7 months ago

          and the next time something writes to that area the data that was there before is disregarded.

          A single overwrite might not be enough to defeat physical forensics because shadows of the old data persist in how the new data is stored. Also when it comes to SSDs you might be waiting a long time for the data to get overwritten as the drive will wear-level its erm sectors (what are those things called with SSDs?).

        • mojo_raisin@lemmy.world
          link
          fedilink
          English
          arrow-up
          9
          arrow-down
          28
          ·
          7 months ago

          So are you saying that they suffered from a filesystem bug that caused deletion failure? I’d imagine they use standard filesystems on their backend, I haven’t heard about any bugs like this.

          If you ask me, what’s more likely, that a company known for shitty behavior lies about deleting files so they can continue to use that information to profit, – OR – that they are experiencing a filesystem bug on their backend, I’ll choose the former.

          • Simon Müller@sopuli.xyz
            link
            fedilink
            English
            arrow-up
            31
            arrow-down
            3
            ·
            7 months ago

            no I don’t believe a damn word of what apple’s gonna say on this, I just wanted to get the message out there that generally file deletion works by allowing data to be overwritten, so if the images are local this could very well just be that either it’s showing data that hasn’t been overwritten yet or it accidentally brought things out of the “recently deleted” depending on how long ago it was deleted.

          • brbposting@sh.itjust.works
            link
            fedilink
            English
            arrow-up
            11
            arrow-down
            2
            ·
            7 months ago

            Undeleting nudes

            That’s iPhone

            Seriously: I don’t think the cost benefit is there to intentionally make a maneuver like this. Any crap they pull needs to have a perfectly proper explanation, with our agreement to a specific term buried somewhere in their policies. Can only imagine how much money they blew throwing these billboards up all over the San Francisco Bay area. We have to buy Apple over Google for ostensible privacy gains, and Apple has to lock us in to their walled gardens to make up for their comparatively smaller ad/data business.

            This post assumes Apple is aethical (that’s like amoral but for ethics right?) but still a self-interested economic actor. They can’t let short-term greed get in the way of long-term greed!

            • mojo_raisin@lemmy.world
              link
              fedilink
              English
              arrow-up
              4
              ·
              7 months ago

              Seriously: I don’t think the cost benefit is there to intentionally make a maneuver like this.

              You might be right

              They can’t let short-term greed get in the way of long-term greed!

              lol

    • lurch (he/him)@sh.itjust.works
      link
      fedilink
      English
      arrow-up
      13
      ·
      7 months ago

      the shred command in Linux tries to do this, but it may not work if the hardware moves rewritten data blocks around to mitigate wear.

      • tal@lemmy.today
        link
        fedilink
        English
        arrow-up
        8
        ·
        edit-2
        7 months ago

        shred doesn’t even necessarily work at the OS level. If you use something like ext3 and I assume ext4, normally when you overwrite data in a file, you’re not overwriting data even at the logical level in the block device. Journalling entails that you commit data to somewhere else on the disk, then update the metadata atomically to reference the new data.

        It was more-practical in an era of older filesystems.

    • Forester@yiffit.net
      link
      fedilink
      English
      arrow-up
      18
      arrow-down
      6
      ·
      edit-2
      7 months ago

      Proper deletion should include writing all ones or all zeroes to the block but y’all be lazy as fuck.

      • cm0002@lemmy.world
        link
        fedilink
        English
        arrow-up
        30
        arrow-down
        1
        ·
        edit-2
        7 months ago

        Only necessary on the ol spinning rust, with SSDs not only is it completely unnecessary, but it also burns extra writes.

        Spinny’s store data magnetically on the platter with 1s and 0s, SSDs store data on the NAND as a held charge. If there’s a charge in the block it’s a 1 if there’s no charge it’s a 0.

        With spinny’s, a file gets marked as “deleted” but the residual magnetic 1s and 0s will remain on the platter until eventually overwritten

        With SSDs a file gets marked “deleted” and within no more than a few minutes TRIM comes along and ensures the charge on the NAND is released for that data, there’s no residuals to worry about like with spinny’s and is in fact necessary to ensure decent lifespans.

        • brbposting@sh.itjust.works
          link
          fedilink
          English
          arrow-up
          3
          ·
          7 months ago

          Wow, the SSD can hold the charges perfectly while unplugged for ages? Amazing.

          In a post apocalyptic world where I am in charge of building a storage drive and I’m given all the instructions and fabs, the world is going without storage.

          • davidgro@lemmy.world
            link
            fedilink
            English
            arrow-up
            13
            ·
            7 months ago

            Wow, the SSD can hold the charges perfectly while unplugged for ages? Amazing.

            Yup. Before flash memory, devices like video game cartridges which had game saves actually needed a battery to power the memory holding the saves.

        • Verat@sh.itjust.works
          link
          fedilink
          English
          arrow-up
          2
          ·
          7 months ago

          But wouldn’t TRIM be the deleting he is requesting? Removing the charges would be setting all the bits in that block to the same value.

      • AProfessional@lemmy.world
        link
        fedilink
        English
        arrow-up
        10
        ·
        edit-2
        7 months ago

        That just makes no sense to do, modern storage is write limited. As long as you used encryption the old bits mean nothing to anyone but you.

      • EvilBit@lemmy.world
        link
        fedilink
        English
        arrow-up
        9
        arrow-down
        1
        ·
        edit-2
        7 months ago

        I’m not an expert, but wouldn’t proper deletion be writing random ones and zeroes to the block? Multiple times?

      • Simon Müller@sopuli.xyz
        link
        fedilink
        English
        arrow-up
        3
        ·
        7 months ago

        yeah cuz for normal, day-to-day use that’s exponentially slower the more you’re deleting

        You can do that when you wipe something.

    • solarvector@lemmy.zip
      link
      fedilink
      English
      arrow-up
      4
      ·
      7 months ago

      That’s skipping over the fact that recovering deleted data, even if it isn’t overwritten, is not an “oops”. It it takes extra effort, and if that data isn’t being protected it would be overwritten incidentally as drives are used.

      There is a big difference in a database between “flagging” data and actually removing the association of the data to the database.