Overfitting

What is overfitting? Why is it so problematic for Decision Tree Induction? How to address overfitting?