VISUAL TOOL FOR MESSAGE FLOW ANALYSIS
Original Publication Date: 2002-Nov-29
Included in the Prior Art Database: 2003-Jun-21
Coverage analysis, in general, is used for finding areas of a program that were not exercised by a set of test cases. This leads to creating additional test cases to increase coverage. Coverage analysis can help in monitoring the quality of testing, assist in creating tests for areas that have not been tested before, and help with forming small, yet comprehensive regression suites. For software programs, a large variety of coverage measures exist, such as Statement Coverage, Decision Coverage, Condition Coverage, Multiple Condition Coverage, Path Coverage, Call Coverage, etc. Each has its strengths and weaknesses. A message flow can be viewed as a visual program (composed of a graph of connected processing nodes and the ESQL programs encapsulated in them). Therefore, the coverage analysis concept is applicable to message flows when coverage models are modified to apply to them. The following coverage measurements can be defined for a message flow: Processing Node Coverage – This measure reports whether each processing node was executed. The model can be refined to report whether each port of the processing node was passed through. A coverage task in this context is a single processing node. The set of coverage tasks is the list of all processing nodes in the message flow. For a primitive node, once the node was entered, it is covered. For a compound node, several measures can be defined: Full coverage – if all nodes in the flow were visited; and partial coverage – according to the percentage of nodes in the flow that were visited.