Like everything, the answer is "it depends." I wrote a book a decade or more ago on an F/OSS project I was involved with. The advance worked out to about $2/hour, and I've never sold sufficient copies to go beyond that. That being said, it was always intended more as a labor of love and a way of promoting the project than it was a way to make money. Sadly, there was a political battle within the project's core team, and the fundamental API of the project changed around in a non-backwards-compatible way. That rendered my book semi-obsolete within months of publishing. I left the project shortly thereafter. While those externalities don't bear directly on the book-writing experience, they may be things to keep in mind: stability of the subject you're writing about, general versus specific applicability, and the health of the community you're writing for.