I'm not exactly sure why that is either, but I do think most developers are appealed by the idea that a computer program is a closed system and thus any error that happens is a fluke, unexpected, and should be treated as catastrophic, when that is simply not true. Programs take input from the parent reality, asynchronously, and therefore aren't just subject to cosmic rays flipping bits in the hardware. Personally, I like to handle the possible errors intelligently, but many developers, especially web developers, love to just have any error result in an error page. Even worse, they sprinkle `try { ... } catch (err) { // do nothing }` all over the place, and that results in problems that are hard to track down or detect early.