Knowledge itself can be subdivided into two types: explicit and tacit. We're aware of explicit knowledge: that is, we know we know it. Consequently, we can codify explicit knowledge and communicate it in books, manuals and working procedures, copy these documents, share them or store them in a library or database.
However, because we can codify this knowledge we can get a false sense of security: reading a copy of The C++ Programming Language may look like learning C++, but it leaves much unsaid about the language and its idioms, styles and norms. Expecting someone to maintain a C++ program after reading a book, or even going on a short course, is about as reasonable as expecting someone who has been on a car maintenance course to change the tyres during an F1 pit stop.
 See Stroustrup (1997).
 Thanks to Alan Griffiths for this analogy (private correspondence).
There's a second, more subtle, form of knowledge at work, which is tacit knowledge. This is more difficult to codify and write down. We normally learn it by some process of osmosis. Much of this knowledge is embedded in our work environment, culture or skills, or just the way we work. Often, we don't recognize that we know something special.
The classic example is riding a bike, an everyday thing that most people can do. However, nobody ever learned to ride a bike by reading a book – learning to ride a bike is a process where we learn through experience.
Tacit knowledge can be communicated ...