Elevators

Why your elevator feels personal — and the comments are losing it

TLDR: The article explains that modern elevators use smarter decision-making to reduce long waits, especially during rush hour, instead of simply sending the nearest car. In the comments, readers turned it into a comedy roast about traumatic coding assignments, cursed interviews, and the suspicious feeling elevators are out to get us.

A seemingly innocent explainer about why elevators take forever somehow turned into a full-on group therapy session for people traumatized by bad lifts, bad coding homework, and even worse job interviews. The article walks through the big idea in simple terms: elevators aren’t just going up and down at random. Older systems follow basic patterns, while smarter ones constantly reshuffle which car should answer a call, trying to cut those brutal waits that people actually remember. Not the average wait — the “why have I aged 10 years in this hallway?” wait.

And wow, the community had feelings. One commenter said the post triggered “war flashbacks” to an object-oriented design interview, while another called building elevator control software one of the most frustrating assignments of their entire computer science education. That’s the vibe: half the thread is nerd nostalgia, half is unresolved emotional damage. Then came the funniest hot take of all — the darkly relatable suspicion that the people programming elevators might secretly be “evil sadists.” Honestly? Anyone who’s stood in front of two closed doors while the car skips their floor has considered it.

The other big mood was playful obsession. Multiple commenters immediately linked Elevator Saga, a browser game where you program elevators yourself — basically the thread’s way of saying, “Oh, you think elevator logic is easy? Go ruin your evening with this.” The result is a delightfully dramatic comments section where a basic building convenience became a villain origin story.

Key Points

  • The article explains the SCAN algorithm and the LOOK algorithm as foundational approaches to elevator movement.
  • In multi-car systems, a central scheduler can assign requests to the nearest elevator, but this is presented as a basic rather than optimal strategy.
  • Elevator performance is evaluated through wait-time metrics, including threshold measures and percentile distributions such as p50 and p90.
  • Traffic patterns strongly affect elevator performance, with morning office rush producing especially poor wait-time statistics.
  • Otis’ RSR algorithm improves dispatching by scoring cars with multiple factors and re-optimizing assignments every five seconds.

Hottest takes

"war flashbacks about a particularly irritating OO design interview" — shadeslayer_
"the people who apply the logic ... are actually evil sadists" — orliesaurus
"One of the most frustrating assignments ... in my computer science education" — clarkmoody
Made with <3 by @siedrix and @shesho from CDMX. Powered by Forge&Hive.