Linear superposition based on simple theory suggests that the improvement in linear bit density for the thin-film head is about 25% over its Karlquist counterpart. Dynamic iterative simulation via the Extended Dynamic Iterative Model (EDIM) concluded that a finite pole-width head is excellent for writing but poor as a read head due to limited low-frequency response,and the previous superposition prediction appears optimistic. A good combination would be a narrow polewidth for WRITE and a wide pole-width for READ. The criteria of linear and non-linear distortion for magnetic recording are reviewed, and a group of universal test patterns, capable of revealing worst-case peakshift and resolution independent of coding schemes, is proposed. The modelling technique of EDIM is briefly discussed with illustrations of the recording and demagnetizing distribution for in-depth analysis of the high-density digital recording process.