During task composition, such as can be found in§distributed query processing, workflow systems and AI§planning, decisions have to be made by the system and§possibly by users with respect to how a given problem§should be solved. Although there is often more than§one correct way of solving a given problem, these§multiple solutions do not necessarily lead to the§same result. Some researchers are addressing this§problem by providing data provenance information.§Others use expert advice encoded in a supporting§knowledge-base. However, users do not usually trust§complete automation during decision-making for§certain domains with natural variation, like biology;§they need a way to be able to control and/or§intervene with the system's reasoning to verify parts§of the process. This book provides a thorough§analysis of the problem and presents a data-centric§methodology of measuring decision criticality and§describe its potential use. We argue that agent§technology is a natural fit for the design of§distributed heterogeneous integration systems,§particularly in bioinformatics, and we propose a§multi-agent system design and architecture as the§basis of our framework.