The overview of MOLGEN begins with a definition of problem solving. It states that it is "the process of developing a sequence of actions to achieve a goal." Many of the systems reviewed so far do not necessarily develop this sequence of actions, but rather carry out the actions and are still considered problem solvers. Granted, once the actions are known, many problem solving tasks are trivial in comparison. The work and intelligence needed in carrying out the actions can still be significant.

      I found difficulty in distinguishing the differences between hierarchical planning and non-hierarchical planning on a practical/system/implementation level. While the concept is clear, a diagram such as that provided for illustrating a hierarchical structure of sub-plans for non-hierarchical planning would have been very helpful.

      An example the author used often in explaining interacting subproblems is that of painting both a ladder and ceiling. The basic premise is that the ceiling must be painted before the ladder because one cannot stand on a freshly painted ladder. It seems ill suited since a plausible solution could be paint the ladder, wait for it to dry and then paint the ceiling. Without a time constraint or with the lack of a "wait" function/component the example doesn't serve its purpose very well.

      The author addresses any surprise the reader may have in regard to the simplicity of the skeletal plan refinement method. We are reminded that it attempts to produce competent plans, not innovative ones. It should also be noted that the simplicity of the skeletal approach becomes extremely valuable as the number of skeletal plans and options within each increases. One would be hard pressed today to choose from the variety of proven methods available without assistance from several domain experts or an expert system.