This paper summarizes the literature on research studies in informatics and changes in learning, particularly computer-assisted instruction. Research results themselves are often inconclusive when no allowances are made for different generations of hardware and software or computer types, when the computer is treated as the only variable in the scenario, or when the epistemological roots of classroom practice are not acknowledged. Contradictory commentary springs from the fundamental dichotomy between the mechanistic and constructivist approaches. Their vastly different perspectives create disagreement over whether classroom computers actually affect cognition, and for how long. (Contains 31 references.) (BEW)