The pervasive association project aims to knowledge process the information about the real world. To achieve this goal, bridging of the human knowledge and the sensor data btained from the real world plays a key role. To tackle this task, we propose a real world event labelling method based on primitive elements of events. In this paper, we focus on selection of the primitive elements based on the notion of motion study and function analysis. By describing concepts that form the human knowledge as a combinations and sequences of primitive elements, the existing knowledge is extended for the correspondence to sensor data.

For the decision of software function requirements, the utilization of information acquired from past developments is the key for the proper and effective achievement. We proposed the methodology for that, which store the knowledge in advance as 'ontology' and help developers utilize those 'ontology' when they develop new software system. We also implement the prototype system and evaluate the proposed process using it.

The composition of Semantic Web Services is to synthesize invoking sequences of Web Services with data flows, so that the sequence of services turns out goals demanded by users. The problem of Web Service composition could be deemed a planning problem on the domain of Web Services, and then the web composer could be a sort of planner. However, the web composer's domain model and the details of realization are, in practice, quite different from ones of traditional planners. The web composer unify neither terms nor literals in preconditions to ones in effects, which include instances, but unify types of input parameters and output parameters of Web Services, which include classes. Therefore, the unification process in the web composer must infer the type matching upon input and output parameters through class subsumption. Furthermore, the dataflow between IO parameters and preconditions and effects makes the problem quite complex. In this paper, we formalize a partial-order planner for composing Semantic Web Services, based on traditional partial-order planner UCPOP, and then discuss the complexity in Web Service composition that involves inputs, outputs, preconditions, and effects. The algorithm of UCPOP is modified according to the simplest case model of Web Service composition. This planner can compose atomic services into a composite process from scratch. The resolution of threat from effects and the mixture initiative of planning will be implemented in the next version.