Building a Custom Rule Engine With Prolog

摘要

Unlike factual or procedural knowledge, logical knowledge is difficult to encode in a computer. However, the ability for organizations to successfully encode their logical knowledge can lead to better services for users. The question, then, is how best to encode logical knowledge. It can be shoe-horned into data- and procedure-based tools, but the encoding is difficult, the knowledge becomes opaque, and maintenance becomes a nightmare. Rule-based and logic-based tools are better suited to the encoding of logical knowledge, but require the selection of the proper tool for the knowledge to be encoded, and the learning of how to use that tool. Off-the-shelf rule or logic tools sometimes provide a good solution, but often the knowledge representation of the tool doesn't fit the actual knowledge, or the reasoning engine doesn't use the knowledge as it is supposed to be used. This leads to the same coding and maintenance problems experienced with conventional tools, but to a lesser extent depending on how big the semantic gap is between the knowledge and the tool. A viable alternative is the building of a custom logic-based language and reasoning engine. This allows for the closest fit between the coding of the knowledge and the actual knowledge, and for the cleanest integration between the tool and the rest of the application context.