A paper published by the author in 1988 in this journal provided some important findings about the lack of precision of visual acuity (VA) measures made with commonly used Snellen charts and the advantages of using letter charts designed using the principles proposed by Bailey and Lovie in 1976. That 1988 paper has been cited a number of times since, mostly supporting the findings. The purpose of this review is to examine the changes that have occurred in VA measurement in research and clinical practice since that earlier study. While precise measures of VA using Bailey-Lovie or ETDRS charts are now commonly used in major, multi-centre research studies, it is disappointing to see that many research papers still report VA measured with Snellen charts and even use Snellen fractions, invalidly converted to logMAR notation, in parametric analyses of VA. Many studies have examined the test-retest variability (TRV) of VA measures in groups and individuals, but it is difficult to determine if clinicians or researchers determine patients' individual TRVs to more accurately detect real changes in VA for each individual. This paper summarises the findings of the 1988 study: (1) Snellen charts and VA notations are not appropriate for accurate clinical and research measures of VA; (2) Charts employing the Bailey-Lovie design principles should be used to provide precise measures of VA. (3) Test-retest variability should be used to determine the limit for detecting significant change in VA. This author suggests that it is time for Snellen charts, Snellen fractions and decimal notation to be confined to the teaching of the history of VA measurement. A request is also made to stop the use of the redundant term 'best corrected' VA (BCVA). Recommended procedures are given for precise measures of VA and accurately monitoring changes in VA in clinical practice and research.