[spdustin]

Agreed that the issue opener is a bit of a troll (and antirez: don't feed the trolls!), but let's step back a moment and separate the content from the subtext.

Context: Error messages reach users, by their nature, when something has gone wrong. It's likely that, in that moment, the user's amygdala has started sounding an alarm that the brain's prediction about the outcome of an operation did not come true. It is, quite literally, an emotional moment. A feeling of vulnerability, even if only for an instant, is the first and most profound response when a user encounters an error. Humans further tend to anthropomorphize the systems they interact with - a natural shortcut the brain takes to understand a system that seems to have intelligence - and a cold, unforgiving error message lashing out in a moment of vulnerability seems almost malevolent.

Content: If error messages considered the momentary mental state of a user when they're feeling raw from their expectation not being met - the error resulting from their action - they would be more helpful to the user, and allow them to be self-sufficient, which increases their feeling of well-being and satisfaction with a system. As developers, we can "be right" or we can "be happy" - sure, it's right to say something went wrong. But it's happy when the friendliness of an error message means that a user can resolve an error condition without having to submit an issue to the help desk.