An inter-action diagram is a graphical view of computation processes and communication between different entities. It can be used for the design and testing of distributed systems. In particular, interaction diagrams offer significant advantages to the design of multi-agent systems, especially when they can be expressed in a linear form, known as fragmentation, facilitating automation of design and testing of such systems. Existing interaction diagram formalisms lack the capability of describing flexible temporal order constraints. They only support rigid temporal order, and hence have limited semantic expressiveness. In this paper, we propose an improved interaction diagram formalism in which more flexible temporal constraints can be expressed