Remote viewing is the supposed faculty which enables a percipient, sited in a closed room, to describe the perceptions of a remote agent visiting an unknown target site. To provide convincing demonstration of such a faculty poses a range of experimental and practical problems, especially if feedback to the percipient is allowed after each trial. The precautions needed are elaborate and troublesome; many potential loopholes have to be plugged and there will be strong temptations to relax standards, requiring exceptional discipline and dedication by the experimenters. Most reports of remote viewing experiments are rather superficial and do not permit assessment of the experimental procedures with confidence; in many cases there is clear evidence of particular loopholes left unclosed. Any serious appraisal of the evidence would have to go beyond the reports. Meanwhile the published evidence is far from compelling, and certainly insufficient to justify overthrow of well-established scientific principles.