Setting this into an emotion-theoretic context, let us suppose that Steve has the following three goals:
Goal: (Serendipity) Steve wants the user to be familiar with where the dipstick is and how to remove it. Simulator messages meet these constraints: (a) Show that the user is interacting with the system. (b) Show that the dipstick has been pulled out. Certainty: 100%, based on direct check of ``hard'' system events (see discussion of emotion intensity variables, below). Goal: (preservation) Steve does not want the user to quit the tutoring session without being familiar with where the dipstick is, and how to remove it. Simulator messages meet these constraints: (a) User is quitting a tutoring session. (b) Simulator message history shows that the user has not removed the dipstick. Certainty: 100%, based on history of actions user has performed during the tutoring session. Goal: The user should check the oil level before starting the compressor. Simulator messages meet the first constraint following, and some configuration of the latter constraints: (a) show that the user has started the compressor. (b) Message history shows an instance of the following situation: (1) Dipstick is pulled out, and Steve did not do it. (2) Dipstick is in the user's field of vision when Steve pulls out, and replaces, the dipstick. (3) It is reasonable to infer that the user was looking at the dipstick. Certainty: calculated as a function of context, and perceived sensing on the part of the user.
The first goal above is a serendipity goal, wherein Steve is happy if the situation comes about but not distressed if it does not. The second is a preservation goal, wherein Steve may be distressed if the situation comes about, but is not happy if it does. In both cases, certainty will be 100% (the default value), and based on a direct examination of the messages reflecting user interactions with the ``hard'' virtual world.
With the third (bi-valenced) goal Steve may be happy if the user is perceived to have checked the dipstick before starting the compressor, but distressed if the user is perceived to have started the compressor without checking the oil level. In this case certainty will depend on the level of belief Steve has about whether or not the user has sensed the oil level on the dipstick.