A slim, accessible book on part of this question (and only part!) is Decision Theory and Rationality by José Luis Bermúdez (Oxford University Press 2009). It requires little or no technical knowledge of decision theory, and shows how decision theory can't possibly be an exhaustive account or explication of rationality. Bermúdez makes a good case, in simplest terms, that rationality plays at least three key roles: the guidance of action (i.e. answering the question what counts as a rational solution to a decision problem), normative judgement (answering the question whether a decision problem was set up in a way that reflected the situation it is addressing), and explanation (answering the question how rational actors behave and why). He argues that no form of decision theory (there are lots, and he only explores a few of the more common ones) can perform all three of these roles, yet that if rationality has one of these three roles or dimensions, it has to have all three of them. So decision...

As useless as art, literature, music, or just about anything you do as an end in itself rather than a means toward some other end in itself. Most important science wasn't done for the purpose of achieving "practical" results, but to satisfy some inner compulsion, of the kind that Plato describes very well in his cave story. That said, just about all scientific disciplines have emerged, one way or another, from philosophy (though philosophy itself, if you consider Plato as an important milestone, seems to have been inspired by mathematics), so it's certainly been of very considerable practical use in that sense.