Abstract
Discussion of ``Least angle regression'' by Efron et al. [math.ST/0406456]