This paper presents an approach for interaction between a human user and a mental model-based spatio-temporal assistance system based on the definitions for conceptual and inferential cognitive adequacy (cf. Knauff et. al., 1995). The paper introduces work in progress that covers ideas towards a visual interaction language that provides cognitively adequate knowledge representation and interaction during cognitively demanding spatio-temporal tasks. A trade fair example described in the paper provides an application domain for spatio-temporal problem solving and reasoning tasks. Ideas for adequate knowledge representation, interaction, and reasoning with external media are discussed.