Counterexample: if you know anything about cryptography -- anything at all -- then Dan Brown's earlier book Digital Fortress is downright painful.

Digital Fortress is so awful, I couldn't even get through the first 10 pages.

Err yeh. In retrospect Dan brown isn't as good an example as I'd hoped :-)

Try Robert Heinlein; I learned a lot from the books he wrote in the 1950s. And he actually cared about technical details, as far as the science of the time allowed; he spent hours calculating orbits just to get the timing in Space Cadet right.

Come to think of it, I read a lot of fiction and sub-consciously engage in filtering out the fictional content from the real-world facts in the text.

I'm inclined to believe, that this has helped me become a better filter for the noise/B.S from the real-world too. :)

what earlier books? I lived in rome for about 20 years and trust me, "angels & daemons" is way far from 'researched' even at the obvious geographical details.

Though I agree his works do have a lot of crumbs of informations to put someone on a long wikipedia spree :)

Hmm, yes, geographical was perhaps a bit of a stretch for Brown novels. But I did learn a lot about the Vatican, for example, from that novel.