Microsoft Word – 20172018_sem_exam_feedback_main_r001.docx
2017-2018 Software Engineering (M) – Exam Feedback
This general feedback about the SE (M) exam should give you some insights into how the class
performed in the exam as a whole. Overall, the average score was relatively high, with a large
number of students achieving A grades. The rationales and justifications provided showed a good
level of understanding and reflection on the software process. It was good to see that the class
achieved a high level of performance on this exam and were able to show a good knowledge of
software engineering.
Learning Outcomes
1. Develop clear, concise, and sufficiently formal documentation of extensions to
an existing system, based on the true needs of users and other stakeholders;
Coursework only
2. Apply design principles and patterns while designing and implementing simple
systems, based on reusable technology;
Q1,Q2,Q3b
3. Produce documentation appropriate for programs developed in practical
exercises;
Coursework Only
4. Create UML class diagrams which model aspects of the domain and the
software architecture;
Q1a,Q1c,
Coursework
5. Carry out testing of programs and apply simple measurement techniques to
software;
Q3b, Q3c,Q3d
6. Discuss the breadth of software engineering practices. Q2a,Q3a
Question 1: Modelling and Design Patterns
Most students did well on this question. Especially the two UMLs, most student get full or close to
full marks for them. Some students failed to emphasise that no loopholes can lead to illegal
modifications. Most students get full or close to full marks for their java source code for the
Technician class.
Question 2: Design Principles
Most student generally did well on this question. Most students provide well formed arguments
about the cohesion although a few students misunderstood the cohesion type considered (i.e.
functional, not sequential or otherwise or did not provide any arguments to support their
conclusion. Most students correctly identified examples of five types of coupling (Q2c), and generally
provided good suggestions for how to correct two of them. Minor reductions were applied for not
fully outlining the characteristics of the individual coupling or not specifically mentioning how it
could be reduced in this case (with reference to the code). A subset of students confused data,
stamp and content coupling.
Question 3: System Architecture and Testing
Students generally did well in explaining “test-driven development” (Q1). There were some
uncertainties with the system architecture question; the presented diagram represents a basic
multilayer architecture (with an association); while the system could be implemented using specific
versions of MVP or client-server these system architecture patterns were not implied in the design
(e.g. the TCP/IP communication was not indicated); partly credit was awarded depending on the
validation of the arguments.
Some students struggled with questions c). Some failed to identify meaningful equivalent classes for
the test. Some wrote test code that are disconnected with the provided UML and method signature.
Most students correctly identify that the test should be a black box test.