Reference: Keller, R. M. Operationality and Generality in Explanation-Based Learning: Separate dimensions or opposite endpoints? Stanford, CA, 1988.
Abstract: One view of explanation-based learning (EBL) systems is that they take a non- operational description of a given target concept, and transform it into an operational description. Current EBL systems always specialize the target concept in the process of producing an operational description. This has led to the popular misconception that less general descriptions are necessarily more operational, and inversely, that more general descriptions are less operational. Actually, this particular inverse relationship between operationality and generality is based on a specific set of assumptions that will not be valid in all EBL situations. Therefore, any EBL maechanism that equates "more operational" with "more specific" will work properly only for a subclass of problems. This paper sets forth the position that operationality and genrality are more appropriately viewed as two seperate, potentially independent dimensions, rather than as opposite endpoints on a single dimension.