Decomposition of data flow diagrams

Data flow diagrams are an important design aid in system development. CASE tools allow data flow diagram construction and modification to be automated. Decomposition is the top-down development of a data flow diagram starting with the system inputs and the system outputs. Decomposition may also be automated, resulting in an interactive process for data flow diagram design. Adler (1988) described an algebra for the decomposition of data flow diagrams. A set of quality measures was also described. The authors show that these quality measures do not correspond to the intuitive notion of a good decomposition. A new set of criteria is proposed which does correspond to the intuitive notion of a good decomposition. The use Adler´s algebra leads to an inefficient decomposition process, as well as one which is not guaranteed to find a good decomposition. The authors give an efficient algorithm which gives a good decomposition