The classical path method proposed by Neilsen and Gordon [J. Chem. Phys. 58, 4131 (1973)] is tested by comparison with extensive fully quantal calculations and recent measurements of cross-sections for HF-Ar. Widths and shifts are found to be in satisfactory agreement both with the close-coupling and experimental results. Comparisons are also made with more "downgraded" theoretical approaches based on the peaking approximation and the neglect of the time ordering. Finally, the Neilsen and Gordon approach is used in order to demonstrate that vibration-rotation coupling significantly contributes to the rotational dependence of both widths and shifts cross-sections.