CpSc 372

Exam 2-Key

 

Name______________________________________________

 

Answer any FOUR of the five questions. Place a large X on the question you are omitting. If you answer all five, the one on which you score the best will NOT be counted. All questions are counted the same: 25 points each. Answer all parts of each question.

 

You may have the class notes, handouts, copies of submitted homework. You have from 12:30 pm until 1:45pm for the exam.


  1. Design and architecture
    1. Designing distributed systems involves multiple machines and perhaps multiple machine types. How does the middleware bus design pattern address the issue of different machine types?

 

The middleware pattern uses objects at each end of the data flow to modify the data to fit the format expected by the destination machine.

 

 

    1. Identify one design pattern you used in your traffic intersection design. Describe the forces that led to using that pattern and describe the forces that were still present after the pattern was used.

 

Many different answers possible here.

 

    1. Describe the forces that lead to the use of the callback design pattern.

 

The forces of efficiency and flexibility lead to the callback pattern. When one thread can do work while waiting for another thread to send it data callback objects can be used to wait for the data and then transfer the data to its parent object. In an application framework standard interfaces for callback objects are specified so that the user of the framework can define a callback object that will wait to be called by the main thread.

  1. Threads are the main units of work in a software system.
    1. There are several schemes used to avoid deadlock including at least one design pattern. What is the basic idea that all of them share? In other words, how do you design to avoid deadlock?

 

Avoid unprotected shared data

 

    1. Under what circumstances is it possible to have two threads in the Running state at the same physical time?

 

Have two or more physically distinct processors

 

    1. One element in the integration test fault model is the interaction of threads in the elements being integrated. Describe the basic idea for test cases that could be used to detect potential interactions?

 

Test cases of methods in one object that invoke methods in the interface of the other object will detect certain types of interactions.

 


  1. Below is a state diagram for a simple traffic intersection.

 

    1. Describe the trigger, guard, and action that would be associated with the transition from Green to Red in the TrafficSignal model.

 

Guard: check that light is Green

Trigger: timer or detector sends an event

Action: set light to Red

 

 

    1. The state diagram describes the dynamic behavior of a design element while the class diagram describes the static structure of a design element. What are the possible correspondences between a state diagram and the class representation for a single design element in a class diagram?

 

A method in a class diagram may be the guard, trigger, or action on a transition. An attribute may be data that is checked by the guard or set by the action of the transition.

 

    1. Describe completely two test cases that could be defined given the information in the diagram.

 

(in DominantStreetGreen, GlobalTrigger, in DominantStreetRed)

(in DominantStreetRed, GlobalTrigger, in DominantStreetGreen)

 

 

 

  1. We have used the PSK to prototype the traffic intersection. The PSK provided several tools such as a compiler and a model checker.
    1. The compiler finds mismatches between the definitions of components by examining the syntactic structure of the components. What does testing find?

 

Testing finds differences between expected (specified) and actual computations.

  

    1. The model checker verifies the truth of propositions such as “It is always true that a light being green in one direction implies that the light in the crossing direction is red.” How is this different from the information gained from testing? How does the model checker achieve that information?

 

Testing only verifies a single instance of a proposition at a time. The model checker covers the entire state space examining all possible cases.

 

    1. The component model of the PSK defines how to wire together component instances. What conditions must be met for the wiring together of two component instances to be correct? What “travels” over the connection between two component instances?  

 

Method invocations travel across the connection. This may include data in the form of parameters. What one component will provide must match what the other component will provide.

 

  1. We have discussed several process models.
    1. What is the purpose of the Scrum in agile development?

 

The Scrum is a daily meeting of all team members where progress is checked and a plan is made for the next day.

 

    1. Describe the difference in rhythm (our rhythm has been once a week) between an agile process model and an iterative, incremental process model. How does this difference affect the work assignments given to people working on the project?

 

The agile rhythm is faster. Therefore the work assignments are smaller.

 

    1. In a software product line organization there are two main processes: core asset development and product building. What do these two groups exchange? What is the difference in the goals for the two groups?

 

The core asset team provides assets to the product builders and gets feedback in return. The difference is the time horizon and scope of the two groups. The core asset developers must think about all assets over the life time of the product line while the product builder only focuses on one set of requirements in the immediate future.