Chapter 15 - Data Flow Diagrams

Abstract

Data flow diagrams (DFDs) are a well-known technique from structured analysis that is used to represent the decomposition of the system under development (SuD) into processes and data stores. A DFD shows two things: which processes and data stores can exist in the SuD; and which communications among processes, stores, and external entities can exist. The DFD merely represents the fact that these processes and stores can exist, but not when or for how long they exist. It also represents possible communications, but not how many of them take place nor when they take place. This chapter discusses the elements of a DFD. It discusses external entities, flows, stores, and processes. There are three kinds of processes: data processes, control processes, and composite processes. Data processes may be stateful or stateless. A stateless data process is also called a data transformation. Stateless data processes can be triggered by a T prompt. A stateful process can be enabled and disabled by E and D prompts. If a stateful data process receives an E prompt, it performs its initialization before processing its input. This chapter shows how control processes can be specified by state transition tables (STTs) or state transition diagrams (STDs) and that DFDs can be simplified by introducing variables in the behavior description of a control process.