We have known for twenty years that quantum computers would have unique powers for solving certain classes of computational problems. Throughout these twenty years, workers have striven to identify a physical setting in which high-quality qubits can be created and employed in a quantum computing system. Very promising devices have been identified in several different areas of low-temperature electronics, namely in superconductor and in single-electron semiconductor structures (e.g., quantum dots). The specifications of this control system are now well defined, and are daunting. It must deliver very low noise, precisely shaped pulses in the GHz band for qubit gate control; it must deliver interrogating microwave pulses that sense the qubit state, which must be amplified at the quantum-limited level and delivered quickly back to the control system; because of the nature of the fine-grained error correction needed for reliable quantum algorithm operation, subsequent control pulses must be determined by a rapid (classical) calculation performed using the measurement outcomes as inputs.