How can learning science investigate and build upon the work of Lovett's (2001) studies on component skill practice? Can we uncover ways to affect overall task performance disproportionately to the amount of time invested in that task? Under what conditions do students benefit from whole-task practice vs. component-based practice?