My comments were directed as much to us when we acquire expertise to be mindful of how much our listeners need to know (readers, friends, and family all included). Issac's comments do raise the important issue of how should we interview SMEs and then what is our role to our readers as surrogate SMEs?
Components of expertise
OK, let's the get the obvious out of the way: knowledge. JAVA experts know a lot JAVA syntax and stuff. Historians know a lot of events and dates. That's the easy part.
Studies of experts have discovered that experts see patterns that non-experts do not. For example, athletes talk about being able to "see" the court or the field. Part of why Payton Manning can call such effective audibles is that he can see the patterns in the defense (whereas you or I would see 11 people). When I taught electronics, I would like to start the week by showing a typical schematic we would be dealing with that week and ask the students to estimate how many components there were. The answer was typically "hundreds." At the end of the week I'd ask the question again and the answers were more realistic (20-30). What changed? The students now saw the schematic as a power supply, pre-amp stage, and amp-stage. They saw the patterns and that helped them process the previously overwhelming details.
Another thing research has shown about experts is that their knowledge is tacit--they no longer know what they know. They draw on their knowledge so instinctively they cannot observe their processes. I tried to document a couple of my Dobro picking patterns for a friend and it was HARD! Not the transcription and notation part, but just being able to slow down and see what I did instinctively.
So part of our job as communicators is to help SME's uncover their tacit patterns so we can pass those along to our readers. In that way, we start to transfer expertise instead of just information.
I remember once interviewing an expert and asking what a good starting value was for a particular variable. "It doesn't matter" was all he would say. So I finally said, "OK let's start with a million." In about five seconds we arrived at 35 is a good starting value. After that, it was just "When would you make it bigger? When would you make it smaller? How would I know if it were too big or too small?"
And this isn't just about technical writing. We are all SMEs at something. I've started writing music down and it's forced me to investigate the tacit patterns I've been applying. It's made me a better player and will enable me to be a better teacher if I can ever get one of my grandkids to take up an interest in Hootie's hillbilly music :-)