Noting the disappointing results of past experimentation with computer modeling technology in higher education, the author discusses developments which promise potential: communication between model builders and users, interaction between large- and small-scale models, interface with operating data systems, emphasis on outcomes, and continued experimentation. (JT)