A table that indicates actions to be taken under various conditions, the decision being the selection between the alternative actions. Conventionally a decision table has four parts that are named and laid out as shown in Fig. a. The condition stub part lists the individual inputs upon which the decision depends, while the action stub part lists the alternative actions that may be taken. The entry parts then show the conditions under which each action is selected. This is done by arranging the condition entry part into columns, where each column specifies some condition on each of the input values, and then placing a cross in the same column of the action entry part to indicate the particular action to be taken. All the conditions of the column must be satisfied in order for the column to be selected. Normally the complete table covers all possible combinations of input values in such a way that application of the table always selects precisely one action (see also else rule). The example in Fig. b shows a table for deciding how to travel to work. A ‘—’ symbol in the condition entry part indicates ‘don’t care’.
Decision tables have been used both for program specification and implementation, the latter being achieved by directly interpreting (or generating an executable program from) the decision table format.