• Kalkaline
    link
    fedilink
    arrow-up
    48
    arrow-down
    1
    ·
    edit-2
    1 year ago

    God damn, whoever came up with that is clever. I would have never come up with that on my own.

    • acceptable_pumpkin@lemmy.world
      link
      fedilink
      arrow-up
      55
      ·
      edit-2
      1 year ago

      Some security camera systems have this built in. They show snapshots of various times where you choose the total period, say 24 hours. Then you glance through the snapshots that are all displayed at once on the screen and click on the last one where your bike was still there. That will then “zoom in” the timeline and show another set of snapshots, though this time within a smaller total time window. Keep clicking on the last panel with the bike, and it will soon show you the clip of the bike being stolen.

      Really helpful to find out when something changed.

        • justJanne@startrek.website
          link
          fedilink
          arrow-up
          21
          ·
          edit-2
          1 year ago

          If you’ve got 14 billion years, a theft takes a minute, then you need 53 recursion levels of binary search to find the moment of the theft. (14 billion years can be split into about 7.3e15 1-minute segments, 53 levels of binary search allow you to search through 9e15 segments)

          That means OP assumed that it’d take 1 minute to decide whether at a certain still frame the theft had already occured or not, to compute the new offset to seek to, and the time it’d take to actually seek the tape to that point.

          Not an unreasonable assumption, but a very conservative estimate. Assuming the footage is on an HDD and you’ve got an automated system for binary search, I’d actually assume it’d take 5 seconds for each step, meaning finding a 1min theft on 14 billion years of footage would take 5 minutes.

        • Anemia@lemmy.world
          link
          fedilink
          English
          arrow-up
          4
          arrow-down
          1
          ·
          1 year ago

          According to my napkin math it would take longer than an hour if the tape was ~3.3*10^218 sec long (or three million trillion trillion… (18 trillions) …trillion years). Assuming you have only have two options to choose between but can pick which alternative in in 5 seconds (2^720) and you want to get down to a 1 minute intervall.

          So i mean its not impossible to find a tape long enough though it seems unlikely that we would be so off in our estimates of the age of the universe.

    • jmcs@discuss.tchncs.de
      link
      fedilink
      arrow-up
      38
      ·
      1 year ago

      What if you had to guess a number between 0 and 100 and the other person (or an application) only told you if the number is bigger or smaller? That’s the form that’s usually presented to CS students and most people end up figuring it out on their own. Then the trick is knowing how to generalize it.

    • systemglitch@lemmy.world
      link
      fedilink
      arrow-up
      18
      ·
      1 year ago

      Honestly you probably do it already without thinking about it when trying to figure out where you left off a video that you never paused.

      Or if you ever had VHS tapes, or so e from of disc media… perhaps a cassette when looking for a particular part of a song.

      Maybe not as methodical as perfectly breaking it down into halves of halves, but xlos enough to help you pin point what part you are looking for.

        • Kalkaline
          link
          fedilink
          arrow-up
          11
          ·
          1 year ago

          I’m pretty sure I was serious. I don’t know how people can be that clever. It seems simple once it’s explained, sure, but I wouldn’t be able to come up with that on my own without someone else giving me a problem that points me in that direction.

          • khannie@lemmy.world
            link
            fedilink
            English
            arrow-up
            6
            ·
            1 year ago

            Studied this in computer science algorithms class waaaaayyy back in 1996 and by golly this one stuck with me. It’s so simple and so effective.