DataLex inferencing software logic and assumptions

Key logic and assumptions of the DataLex inferencing software include:

• Inferencing proceeds by both backward and forward chaining whereby the system runs all rules it can without asking a question so that all facts that are knowable given a set of facts are known.
• When an attribute being evaluated is the conclusion of more than one rule in a knowledge-base, the rule set is evaluated top-down, so that the top-most rule with that conclusion is evaluated first. (This can be changed via the ORDER statement.)
• Rules are evaluated strictly from left to right. For example, if a rule contains the expression ‘A or B’, DataLex (ysh) will always evaluate whether A is true, even if it is already known that B is true.

Further information on the logic and assumptions underlying use of the DataLex software can be found in AustLII's DataLex Developer's Manual (June 2019).