This paper provides a review of the present status of the topic dealt with. The contact and non-contact methods used for rock joint roughness measurement are summarized including their salient features, advantages, and disadvantages. A critical review is given of the empirical, statistical, and fractal-based methods used for rock joint roughness quantification identifying their salient features, shortcomings, and strong attributes. The surface topography of rough rock joints is highly erratic. Fractional geometry is better suited than Euclidean geometry in representing highly erratic rock joint surfaces. The influence of non-stationarity on accurate quantification of roughness is discussed. The existence of heterogeneity of natural rock joint roughness and its effect on computed roughness parameters are well illustrated. The controversial findings that have been appearing in the literature on roughness scale effects during the last 40 years have resulted from neglecting the effect of roughness heterogeneity on scale effects. The roughness heterogeneity controls the rock joint roughness scale effect, and it can be either negative, positive, or no scale effect depending on the type and level of the roughness heterogeneity of the rock joint surface. The importance of consideration of the existence of possible anisotropy in the quantification of roughness is well illustrated. The indices available to quantify the level of anisotropy are given. Effects of sampling interval and measurement resolution on the accurate quantification of roughness are discussed. A comparison of results obtained by using different quantification methods is discussed. A few recommendations are given for future research to address the shortcomings that exist on the topic dealt with in the paper.