Feb 25, 2000 David DiVincenzo lists the requirements for building a quantum computer

By the end of the decade there were several ways to make a qubit: trapped ions, nuclei in molecules, superconducting circuits, and proposals based on single electrons in semiconductors. Each laboratory could report progress, but on its own terms, and the results were hard to compare. What, exactly, would a technology have to achieve to become a quantum computer?

In February 2000 David DiVincenzo, a physicist at IBM, set out an answer that became the field's common yardstick.

His requirements were five. The system must hold well-defined qubits and be able to grow to many of them. Every qubit must be set to a known starting state. The qubits must keep their quantum state far longer than an operation takes. There must be a universal set of gates, such as the controlled-NOT together with single-qubit operations. And each qubit must be readable at the end. Two further requirements concerned sending quantum information from one machine to another.

The list did not say which technology would win; it said what any of them would have to deliver. From then on, a claim to have built a quantum computer could be tested against it, point by point.

the requirements for the physical implementation of quantum computation are discussed. These five requirements, plus two relating to the communication of quantum information, are extensively explored