"She's not troubled, she's a dancer"
- Sir. Ken Robinson on a child whose parents feared she might be learning-disabled, thus demonstrating that different types of intelligence exist. Look up the speech on TED Talks; it is incredible.
From my personal life, my wife couldn't care less about computers and probably would never watch a TED video, but her "people intelligence" is very high -- she can relate to many different kinds of people and wherever she goes she makes good friends and people around her hold her in high esteem.
Her dad is something of a bookworm who knows a little bit about everything and quite a bit about a couple of things. If he were younger (he's in his 70's) he'd probably get his news from HN. I've often wondered if his "intellectual intelligence" was transferred to her but in a different form... like she used it in a different way, in interacting with and reading people, instead of books.
Coders can often walk though large logic problems instantly, but IMO that's the same basic thing as an artist knowing which 12 lines capture the basic idea of a picture. Think of it like the brain missing the correct API's, you can fake it but it's slow and buggy. But, if you start to spend 10,000+ hours solving the same types of problems and you are going to create that foundation.
EX: Cab drivers develop much better spatial reasoning skills over time. http://news.bbc.co.uk/2/hi/677048.stm
Trying to explain that to a non-coder is not easy to do without appearing arrogant, and I think that is why people in IT can appear rude. How do you politely say, "Your mental representations are so messy and imprecise I'm having trouble categorizing what you need?"
You say: "OK, now that we've covered the standard flow, let's talk about the strange cases. What are the weird situations that come up once in a while?"
For many users, "ALWAYS" translates to "always, except for the oddball case", and "NEVER" translates to "not usually, but you never know."
That's why Requirements Analysis is as much about psychology and anthropology as it is about logic.
The number of real world problems where ALWAYS applies is vanishingly small. This is the reason that rule based artificial intelligence floundered. The answer to almost every interesting question is some variation of "sometimes", "often", "almost always", "except when", etc. This is why statistical techniques and machine learning have led to solutions to many problems where rule based AI was an abject failure.
http://www.ted.com/talks/ken_robinson_says_schools_kill_crea...