Design methods in rock engineering are well established for the conditions commonly encountered in civil and mining engineering. Empirical approaches are applicable, as well as the use of available rock mass failure criteria, and unknowns can be dealt with by applying an increased Factor of Safety. However, in the deeper, more competent rock mass conditions being encountered in mining and tunnelling, the design approaches may be inappropriate and prediction of rock mass and rock material behaviours is required. This is much more difficult than design and a key aspect is the identification of the mechanisms of rock and rock mass failure. The mechanisms involved in the development of failure are fundamental for the prediction of behaviour. Failure may initiate by one mechanism, propagate by another one or more, and may finally be due to yet another. This paper deals with the issues of design and prediction in rock engineering. Since extensile fracturing is commonly the mechanism by which failure initiates, this mechanism is a focus in the paper. Examples of fracture and failure development in hard rock are presented to illustrate the issues involved in prediction of rock engineering behaviour.