In this paper, reduction and its pragmatics are discussed in light of the development in computer science of languages to describe processes. The design of higher-level description languages within computer science has had the aim of allowing for description of the dynamics of processes in the (physical) world on a higher level avoiding all (physical) details of these processes. The higher description levels developed have dramatically increased the complexity of applications that came within reach. The pragmatic attitude of a (scientific) practitioner in this area has become inherently anti-reductionist, but based on well-established reduction relations. The paper discusses how this perspective can be related to reduction in general, and to other domains where description of dynamics plays a main role, in particular, biological and cognitive domains.