One of the biggest problem areas in relation to Computer Assisted Pronunciation Training (CAPT) applications, designed for Second Language Learning (SLL), has been that of feedback. As with human teachers, the ultimate goal of CAPT is to provide instantaneous, individualized feedback to the user on the overall quality of the pronunciation made. Attempts at accomplishing this goal have seen the light of day in many different ways. Some applications provide intonation curves and some provide spectrograms, color-coding or numerical scores. Unfortunately these feedback modes rarely inform the user of what is wrong instead relying on the users to self-correct through trial and error [2]. In this paper, I suggest that feedback should adhere to the four ‘K’s, being: 1) Quantitative 2) Qualitative 3) Comprehensible and 4) Corrective. I further try to illustrate the four ‘K’s by relating them to the TASK-LT program which is a CAPT application currently under development at the Natural Interactive Systems laboratory.