MAX

• Homogeneous knowledge representation

  The representation used in MAX is lframes (logic frames). These structures can be nested in one another, a property which directly enables the use of meta-knowledge. All the knowledge in the system, which includes control knowledge such as inference rules and factual knowledge such as MAX's knowledge of the environment, is encoded as lframes. This form of knowledge representation is what enables MAX to be decarative.

• Knowledge Access Efficiency

  Because the behavioral specifications are explicitly encoded in the knowledge base, the risk is run of creating an inractable architecture. Fortunately, a rule-bsaed encoding of an algorithmic reasoning behavior only produces a constant factor slow down, but it still results in a less efficient knowledge access scheme.

• Declarative knowledge: Glass Box hypothesis

  With the homogeneous representation, all the knowledge in the system is open to inspection. This allows access not only to the factual information of the system, but also to the control knowledge. This is also know as the Glass Box hypothesis. As with all things, this access has its pro's and con's. While it does allow the sharing of information over the system, the system itself is stuck at a given level of expressability, which is determined by the expressability of the representation. This declartive structure is what allows MAX to implement its meta-knowledge.

• Meta-knowledge

  As mentioned above, all knowledge, including control and reasoning knowledge, is represented as lframes. This gives the system the ability to reason and operate on not only its factual knowledge, but its control and reasoning knowledge as well. This property also allows an open, homogeneous, declarative architecture, which is the basic premise behind MAX.