|   | How to Read Call Flow Diagrams | ![]() |
The following symbols are used as elements in call flow diagrams. Each symbol is displayed, with its standard labeling, and an explanation of its use.
|
Call flow diagrams structure a dialog using the elements presented in the call-flow diagram key. Each subdialog, represented using the dialog notation, is itself expanded as a call flow diagram. The diagrams are grounded in prompts, interactions, data retrievals and computations, branching and possible call terminations.
Each prompt, interaction, computation and decision box is elaborated in the call flow tables, which provide prompt wording for initial, timeout, retry and confirmation prompts, grammar descriptions for interaction states, and descriptions of computations and database accesses for the compute states.
The actions described above are connected in a graph by a series of directed arcs representing conditionals. After an action, represented by a node, executes, a choice must be made as to which action happens next. Assuming there are no exceptional cases (see next section), such as commands, hangups or fatal subsystem errors, the conditions on the arcs leaving a node are executed in order until one succeeds. At this point, control is transferred to that node. In the diagrams, the order of evaluation is left to right, unless explicit numbering is provided.
In case there are no conditions that succeed leaving an arc, the subdialog is eited.
Some exceptional circumstances cause control to be transferred at a more global level. These events are: