A survey of the current delineation practices in the various U.S. States and Canadian provinces indicated the importance of a set of quantitative delineation guidelines. According to the survey, no computer assisted methods are used for the curve delineation task. The evaluator responses before and after curve delineation, with the objective to obtain input for the curve delineation tasks, indicate that (1) evaluators have difficulties to agree on the type, number, and location of curve delineation devices, just by looking at a curve while driving with low beams at night; 2) subjective evaluation of an undelineated curve tends to provide a recommended number of devices that is considerably lower than the actual number required when placing four devices within a driver's functional visual field of view; 3) the optimal type, number, and location of the delineation devices should be computed with a set of algorithms implemented in a software package. These findings support the development and use of a knowledge-based, interactive software which computes the delineation layout. The approach and center speed measurements before and after delineation installation were conducted in order to assure that there are no adverse effects on driver behavior due to delineation. The measurements indicate that there is no systematic pattern in speed increase or speed reduction before and after the delineation devices are installed.