The idea behind the feat, first proposed in 1998, is to put a quantum computer into a “superposition”, a state in which it is both running and not running. It is as if you asked Schrödinger's cat to hit "Run".
With the right set-up, the theory suggested, the computer would sometimes get an answer out of the computer even though the program did not run. And now researchers from the University of Illinois at Urbana-Champaign have improved on the original design and built a non-running quantum computer that really works.
They send a photon into a system of mirrors and other optical devices, which included a set of components that run a simple database search by changing the properties of the photon.
The new design includes a quantum trick called the Zeno effect. Repeated measurements stop the photon from entering the actual program, but allow its quantum nature to flirt with the program's components - so it can become gradually altered even though it never actually passes through.