A matcher function is a function that compares observed student input to a predicted value in a unique way. As described in Section 3.2.1, “Input Matching”, CTAT provides a number of basic matcher functions for use in Jess rules. If none of the existing functions perform matching in the desired way, you may want to write a new matcher function.

A matcher function in CTAT has a few characteristics. First, it should return a boolean value (true or false). This is so the (predict) function, which calls it, can itself return a boolean value. Second, the matcher function should consider both the predicted input (specified in the predict function) and student's input, both of which are provided as arguments to the matcher function. The example function below illustrates both of these characteristics:

(deffunction equal-with-precision (?predictedInput ?studentInput)
  (< (abs (- ?predictedInput ?studentInput)) 0.02)
)
...
(predict newTextField DONT-CARE 2 equal-with-precision)

In this function, equal-with-precision, the student's input is first subtracted from the predicted input. If the absolute value of the resulting number is less than .02, the function returns true; otherwise, it returns false. The predict function is then used to see if the rule's predicted value ("2") is equal (within .02) to the student's value.

	Note
	To reference the matcher function, the function must be defined in either the wmeTypes.clp file or production rules file before the matcher is called.

To evaluate the result of a function in a production rule, use the Jess function eval:

(deffunction sum (?a ?b)
  (+ ?a ?b)
)
...
(predict-observable-action ?cell-name "UpdateTable" (eval sum 2 3) )