Figure 10

Network ruleform propagation. To generate new Network rules, existing rules are combined in logically consistent ways. In Step 1, two BioEntity Network rules are selected to potentially perform a deduction, based on the Object of Rule 1 and the Subject of Rule 2 being the same (BioEntity Network rules are shown, but the principle applies to any Network Ruleform). Deduction can only be carried out if there exists a Relationship Chain rule whose Input 1 and Input 2 fields match the Relationship fields of the two Network Rules, as shown in Step 2. The Relationship Chain ruleform defines rules for composing Relationships together. Here, the selected Relationship Chain Rule expresses the situation that if "A is-a B", and "B has a feature-type of C", then A's feature-type is also deduced to be C. If these conditions are met, then the new rule is generated at Step 3, consisting of the Subject of Rule 1, the Result of the chain rule, and the Object of Rule 2. Note that the Distance field of the resulting rule is the sum of the distances of the two input rules. Users can control how Relationships behave in combination with each other by specifying the appropriate rules in the Relationship Chain ruleform. For example, the "is-a" Relationship can be made transitive by creating a Relationship Chain rule where the Input 1, Input 2, and Result fields are all set to "is-a".