Opportunistic planning is considered to some HCI experts as an improvement on successive refinement. Successive refinement is the method of planning in which tasks are divided into subgoals, which are further divided into smaller subgoals. Finally, planning and actions on these refined tasks are done and collated back together. The problem with this type of planning is that it is methodical without any room for backtracking while performing task actions. Opportunistic planning is a method that allows the changing of intermediate goals as one proceeds to a specific goal. The approach uses "cognitive specialists" and "blackboards" to accomplish this planning strategy. Cognitive specialists can be viewed as a representation of subject matter experts that are only needed when their particular talent is necessary. The concept of a blackboard is used as a tool to allow collaboration between specialists. Overall, opportunistic planning is advantageous because it allows for error recovery and more flexible planning analysis. However, the hierarchical nature of successive refinement is more simplistic.
Opportunistic Planning is the foundation of both Direct Manipulation and Display Based Action. Direct Manipulation systems must exhibit certain special characteristics that command based systems do not employ. Continuous feedback and recognition of the actions necessary for that feedback are the major requirements. The cognitive aspects of Direct Manipulation include discovery, exploration, and bridging of the Gulfs of Execution and Evaluation, and the Semantic and Articulatory distances. These gulfs and distances are what separates the user from the interface. In order to close these distances, direct manipulation systems should give the user a feeling of "directness. This means the user should barely be aware of using the system. Advantages of Direct Manipulation include ease of learning and/or teaching, and reduced error rate. One comprehensive disadvantage is the loss of accuracy due to the fact that most Direct Manipulation systems have graphical interfaces and require use of a mouse. There are several instances that can be sited for which typing would be more advantageous.
The Display Based Action is another proposed model in which there is an interleaving between planning and the actions performed. This is in obvious support of opportunistic planning where we know users react based on feedback rather than working out a complete plan before performing actions. Display Based Action focuses on the concept that users actually base their every move on the feedback they receive from a system and not necessary from their extensive knowledge. This notion is supported by several studies where users are observed using a system for which they have a priori knowledge. These studies show that the user actually did not know the system well, but based their actions on feedback from the display. For example, it was found that expert users of a text editor, when tested on their knowledge without system feedback, were actually unable to recall what steps to take in performing certain tasks. When allowed to have feedback these users have no problem performing their editing duties with their favorite text editor. This shows that systems should display the maximum amount of information to the users as possible to aid the user in performing tasks. This exhibits the characteristics Direct Manipulation systems because such systems require continuous feedback.
The above concepts are all interrelated based on one common concept, feedback. The Opportunistic Planning allows for feedback in the planning process. Direct Manipulation systems require continuous feedback and incorporates the Opportunistic approach throughout the design of systems as well as actions performed using the systems. Display Based Action is a concrete example of why feedback is necessary in the planning, design, and usage of systems. If the users are not provided with this feedback, then what do they have as support for retention of information, and error recovery. The above information was presented to an HCI class at Virginia Tech which involved a four day extensive look into Opportunistic Planning, Direct Manipulation, and Display Based Action. There were lectures on the topics with a strong emphasis on the relationship between the three related topics. These lectures included several examples given to aid the listeners in understanding the topic being presented. There was a miniature, class participation exercise to demonstrate the importance of Opportunistic Planning in any system for which humans have to interact. Each student was given an homework assignment to reinforce what was previously discussed, and to stimulate thought for the next session. There was also an extensive exercise to demonstrate the advantages and limitations of Direct Manipulation. Finally, there was a detailed discussion on all the material that was geared to acquire feedback about the topic, as well as to stimulate research interest.