We compare endpoint, via-point and dense path constraints for a single footstep.
DART is frozen; all optimized methods share 48 paired conditions and equal update budgets.
Endpoint and via-point each succeed in 30/48 conditions, versus 21/48 for full UCPC.
Full UCPC has lower mean joint jerk, but this does not establish command success or realism.
This corrected pilot supports a bounded comparison, not universal path-control inferiority.

Jin Kim is currently pursuing the M.S. degree at the Graduate school of Advanced Imaging Science, Multimedia & Film(GSAIM), Chung-Ang University, Seoul, South Korea. Research interests include AI-based 3D human motion generation and control. Current work examines motion generation under text-specified action sequences and durations. Another research direction explores multiscale latent representations and diffusion models for generating overall movement and fine motion details in stages.