Early restorative justice proposals and some of its practices were devised out of general opposition to the use of imprisonment. Initial victim-offender reconciliation cases in Canada and the United States focused on crimes, notably arson and home burglary, that would normally have warranted imprisonment. Generally, however, restorative justice measures have not met adherents' expectations for being an alternative to imprisonment. Although there are numerous reasons for this, this paper focuses on only one reason, i.e., the manner in which restorative justice is implemented in small-scale as well as larger projects. Restorative justice projects must be designed and implemented for target populations of offenders who would normally be incarcerated. This involves analyzing local data to determine what types of offenders are usually incarcerated, as well as the sentencing practices that involve the decisionmaking of local prosecutors, defense attorneys, judges, and probation officers. Restorative justice program designers must also know the personal and criminal justice histories of potential clients. Armed with this information, restorative justice advocates or program staff should routinely interact with prosecutors, defense attorneys, judges, and probation officers to teach them about restorative justice, to apply restorative justice to particular cases, and to obtain judicial approval for such a sanction. The criminal justice personnel whose decisionmaking impacts sentencing must be convinced that restorative justice measures are more cost-effective than incarceration; otherwise, their sentencing decisions will not be altered to place in restorative justice venues offenders whom they used to incarcerate. At best, restorative justice practices, where they exist, divert cases from prosecution or court processing; and even these cases are more likely examples of net widening, since they target offenders who would not usually have been prosecuted or formally processed. 15 references