From RIF
Jump to: navigation, search

Comments as at Jan2010

Therefore, each state of the transition system used to represent a production rule system must keep a memory of the previous states, of the rule instances that where selected and of the action instances that triggered the transition in those states.

It might be more accurate to simply state that that ...must keep a memory of the previous states, including the rule instances that were selected and updates to the working memory. The current wording implies a complexity that is possibly overstating the case! PaulVincent 10:39, 27 January 2010 (UTC)

To avoid confusion between the states of the fact base and the states of the transition system, the latter will be called production rule system states.

A few additional definitions are required before a production rule system state can be defined more precisely: rule instance, for instance, can be understood informally as the result of the substitution of constants for all the rule variables in a rule. However, the most widely used conflict resolution strategies require that the system keeps track of additional information associated to rule instances. In the remainder of this document, rule instance, matching rule instance and action instance will be used as formally defined below.

This seems a bit complicated: again it might be best to simply say:

The state of the fact base (aka working memory) has to be considered alongside the state of the rule transition system (aka rule engine), for which we use the term "production rule system states". A "rule instance" is the result of substitution of a set of values for the rule variables in a rule. The conflict resolution strategies require that the system track additional information associated with rule instances. The following definitions are used in the remainder of the document.

Note that of course the idea of rule instances is itself an abstraction of the processing of a Rete-type rule engine... PaulVincent 10:40, 27 January 2010 (UTC)