I used to teach Computer Graphics for almost 10 years, and I like to believe I became somewhat good at it. I've written some articles about fairly technical topics and the reaction I tend to get is "oh wow, that's the clearest explanation I've read".
I think it boils down to two things: form a mental model of the state of the other person's mental model, and find a path from something they already know to the thing you want them to know.
Start from some basic assumptions, state them explicitly if you need to. From this starting point, build the explanation gradually, step by step, and perhaps most importantly, in a way that makes sense to them.
That's how I structured my CG course. We alternated theory classes (where we'd explore some concept, say, drawing a filled triangle) and a practical class (where the students would implement that). So after the "filled triangle" class, they applied the algorithm to their existing wireframe cube code, and most of the time it looked completely wrong, because the order of the triangles was random.
So I'd start the next class with "Last time your cube looked all wrong. Why?". Students would offer their ideas, and I always tried to find something right in them to build towards the idea I wanted them to arrive at. Then I asked "How can we fix this?", and again, depending on their suggestions, guide the conversation towards the algorithms I wanted to "rediscover" (painter's algorithm and depth buffering).
The key was that the students were very motivated to understand why their cube looked wrong even though they had no bugs in their implementation. They wanted it to look right. And I didn't just walk in and say "open your book on page 159, this is depth buffering" (I didn't use a book at all, in fact), but "reinvented" the algorithm with them, based on their suggestions and some gentle nudging in the right direction.
The worst thing in these circumstances is the opposite: "there's this random thing you need to learn, it will make sense in 5 weeks, I promise".
Sorry for the long-winded answer... I hope it applies to explaining stories and movies, and hey, this is HN ;)
I think it boils down to two things: form a mental model of the state of the other person's mental model, and find a path from something they already know to the thing you want them to know.
Start from some basic assumptions, state them explicitly if you need to. From this starting point, build the explanation gradually, step by step, and perhaps most importantly, in a way that makes sense to them.
That's how I structured my CG course. We alternated theory classes (where we'd explore some concept, say, drawing a filled triangle) and a practical class (where the students would implement that). So after the "filled triangle" class, they applied the algorithm to their existing wireframe cube code, and most of the time it looked completely wrong, because the order of the triangles was random.
So I'd start the next class with "Last time your cube looked all wrong. Why?". Students would offer their ideas, and I always tried to find something right in them to build towards the idea I wanted them to arrive at. Then I asked "How can we fix this?", and again, depending on their suggestions, guide the conversation towards the algorithms I wanted to "rediscover" (painter's algorithm and depth buffering).
The key was that the students were very motivated to understand why their cube looked wrong even though they had no bugs in their implementation. They wanted it to look right. And I didn't just walk in and say "open your book on page 159, this is depth buffering" (I didn't use a book at all, in fact), but "reinvented" the algorithm with them, based on their suggestions and some gentle nudging in the right direction.
The worst thing in these circumstances is the opposite: "there's this random thing you need to learn, it will make sense in 5 weeks, I promise".
Sorry for the long-winded answer... I hope it applies to explaining stories and movies, and hey, this is HN ;)