In addition, the acquisition of sound knowledge about the effects of crime control policies is expensive, but it is much less costly than a prolonged commitment to error. A suitable information system should guide the decisions of police, prosecutors, and judges. The main problem faced by policymakers concerned about crime is how to rearrange the existing incentives to facilitate shared goals and systematic efforts to discover and implement new knowledge about how best to attain those ends. This task is possible. Police departments are much more inclined to experiment now than in the past, although resource cutbacks have greatly increased the costs of experimentation. Citizens have created thousands of anticrime organizations, but interest in the more difficult of these has been hard to sustain. Prosecutors have used computerized information systems and other techniques to focus much more on career criminals, although it is not clear that they have focused on the most appropriate offenders. In addition, legislatures have changed sentencing laws, judges have shown a willingness to rethink the current status of the exclusionary rule, new prisons are being built, and meaningful ways of supervising nondangerous offenders in the community seem to exist. State and local officials and private citizens, rather than the Federal Government, have the main role to play in implementing the needed changes in policies and procedures.