What books would you recommend to anyone working with Semantic Web?

Not only for programmers but also for people who create ontologies.

Bob DuCharme's book Learning SPARQL would have to make this list today.

It's a difficult book, but The Description Logic Handbook helped me overcome my fear of an RDF planet.

It's a collection of review papers about research in systems similar to OWL. I had many questions of the form "Why doesn't OWL do X?" that were answered by this book. I learned new ways to use OWL, and was inspired by projects people have done with predecessors to OWL. Semantic Web For The Working Ontologist would be a pre-requisite for reading it.

"Tom Heath and Christian Bizer (2011) Linked Data: Evolving the Web into a Global Data Space (1st edition). Synthesis Lectures on the Semantic Web: Theory and Technology, 1:1, 1-136. Morgan & Claypool." It is even available as Open Book in HTML, so have a look.

"Thinking on the Web: Berners-Lee, Gödel and Turing" by Peter Alesso is a quite non-technical book that talks a lot about concepts behnd the semantic web. Another book by him that also touches related topics is "Connections: Patterns of Discovery". Both are nice reads and might be good choices for people who first like to understand the concepts and philosophy behind the semantic web.

A nice hands-on book is "Programming the Semantic Web" by Toby Segaran. It includes lots of code and examples and provides you with the needed konwledge to get your hands wet fast.

A fun book with a business focus is Final Jeopardy: Man vs. Machine and the Quest to Know Everything.

You might miss it if you're not careful, but this book talks a lot about the evaluation problem for semantic systems. How do we know that they're "good enough?" What do we do to make something "good enough?" One of the reasons why semantic technology has been stuck in the "commercialization valley of death" is that we often evaluate things improperly. Conventional precision and recall metrics mean very little if the system makes mistakes that destroy user engagement and costs sales. Mike Petit of OpenAmplify talks about the "embarassment test" and so will anybody else who wants to make a living on this new technology.

I really liked Intelligent Database Systems, from 2001. Although it pre-dates many semweb developments, it still has a hefty bunch of content relating to our roots in things like CyC and the pre-RDF ontologies (e.g. I think KIF is in there). It's split into two parts, one which comes at things from a Database perspective, the other coming from an AI perspective. Definitely worth the $10-$13 it's going for these days.

There's also Semantic Web Programming by John Hebeler

Zen and the Art of Motorcycle Maintenance: An Inquiry into Values by Robert Pirzig.