This article describes a large number of points that have to be considered when evaluating computer-assisted-perimeters (CAP). With the ever increasing numbers of commercially available CAP over a large quality-range the choice of a CAP for a certain type of practice becomes very difficult indeed. The authors starting position and present-day manual perimetry are first described. The shortcomings of kinetic perimetry are demonstrated. A differentiation of several types of visual field defects is given. Such a differentiation is indispensable for the evaluation of CAP. Each phase of the visual field examination: detection, limited-and extended assessment is considered. Special attention has been given to examination-strategies. They are the most important part of the software of CAP and usually the least comprehensible part. The existing threshold and suprathreshold-techniques are explained. The author expresses a preference for a threshold-related gradient-adapted suprathreshold detection strategy. Random presentation, stimulus-timing, zero-fear presentations and reliability measurements among many other points are mentioned. The possibilities for an extended assessment and especially the data-base for follow-up examinations are described. The extended (or intermediate) assessment should provide sufficient information on spatial and intensity distribution of the defects. The important matter of the graphical display of CAP-results is elucidated. Statistical considerations may be one of the most promising features of CAP. Repetition of measurements will be necessary for such statistical treatments. The accurate numerical description of defectvolume and its fluctuation, and subsequently the absence or presence of significant change are ultimate goals of statistical programmes. General data on the hardware and software of a particular CAP provide the basis for understanding the results of a clinical evaluation of this CAP. The ultimate answer concerning the usefulness of a CAP can only be given by this clinical evaluation. The criteria and problems of such evaluations are discussed and illustrated with an example. In the addendum a suggestion for a classification of glaucomatous visual field defects is presented. The list of references gives over 160 publications related to CAP.
