Figure 1. Subsequent state after a transition
Events or conditions may cause objects to be in more than one state at the same time. To express this in OSA, more than one state may follow a transition. The meaning is that when the transition finishes, the object enters all the subsequent states. A set of subsequent states is called a subsequent state conjunction. (It is called a conjunction because the object enters state S1 and state S2 and ...) It is represented by a branching arrow with one base, at the transition, and multiple heads, one at each of the subsequent states in the conjunction.
Since a set can have just one member, a subsequent state conjunction may consist of just one state. Therefore strictly speaking what follows a transition is always a subsequent state conjunction although it may be a "conjunction" of a single state.
For example, when it is fifteen minutes before closing, docents cut short any tour they happen to be giving, moves the group toward the exit, and thanks them for coming. Figure 2 shows that if a docent is giving a tour and it is less than fifteen minutes before closing, he stops giving a tour and begins moving the group to the exit and thanking the group for coming.
Figure 2. A conjunction of subsequent states.
Subsequent state conjunctions, then, are one way to begin intra-object concurrency.
From some point of view, a transition may result in the object entering any one of several different subsequent states at random. For example, the docent in our example may be able to show either parrots or hawks at will. This nondeterminacy is represented by multiple arrows from the transition, one to each of the subsequent states. Figure 3 expresses the docent showing either parrots or hawks at random.
Figure 3. Nondeterministic choice of subsequent states.
Why don't we find out why one or another state is entered and add the reason as a condition on multiple transitions and thus eliminate the nondeterminacy? It may be unimportant which choice is taken, so the state net is easier to understand, or the analyst may wish to leave the details for another refinement, or the choice may actually be random. In any case, a transition may lead to a nondeterministic choice between subsequent states.
These expressions may, of course, be combined. Subsequent state conjunctions may be one or more of the choices following a transition. Subsequent states may be joined, adding another thread, rather than just turning on the state. Figure 4 shows that if a docent is giving a tour and it is less than fifteen minutes before closing, he stop giving a tour. At random the docent begins one of
Figure 4. Multiple subsequent state conjunctions.
Go to Prior States Threads