.: topics

The following are a list of suggested MS topics. Most of these could be expanded into a PhD dissertation. This list is constantly updated.

Empirical Evaluation of Synthesis from Use Case Charts

Goal: UCSIM generates well-structured hierarchical finite state machines (HFSMs) from a scenario specification. This automates a key transition in OOAD. The goal of this project is to empirically evaluate the UCSIM tool and its synthesis algorithms. In particular, the research question to answer is whether the HFSMs generated by UCSIM match those that would be produced manually. The evaluation will be carried out by examining a number of existing UML designs and using UCSIM to compare the generated HFSMs to those that were derived by hand. Depending on the abilities of the student, improvements to the UCSIM algorithms may be identified and implemented in UCSIM.

Prerequisite knowledge: UML and OOAD

Difficulty Level: Medium

The Role of Executable Scenarios in Education

Goal: UCSIM generates well-structured hierarchical finite state machines (HFSMs) from a scenario specification. This automates a key transition in OOAD. Furthermore, it allows for simulation/animation of a set of requirements scenarios. This has the potential benefit of trapping requirements errors very early in the software lifecycle. The goal of this project is to further develop the UCSIM tool and empirically evaluate its role in education. In particular, the student will apply UCSIM in the classroom and assess its impact on the learning process for students learning UML and OOAD.

Prerequisite knowledge: UML and OOAD

Difficulty Level: Medium.

Goal: Aspects are cross-cutting concerns that lead to tangled and scattered representations if not handled properly using aspect-based techniques. This project will apply the technique of critical pair analysis to automatically detect undesired interactions between aspects during software modeling. This will build upon work by students Praveen Jayaraman and Ahmed Elkhodary. A language UMLT was developed to define model compositions in a way that critical pair analysis can be used. This project will extend UMLT to the definition of aspects and apply critical pair analysis to detect aspect interactions.

Prerequisite knowledge: UML and OOAD

Difficulty Level: Medium to Difficult.

An incremental code generator for UML statecharts

Goal: Standard code generators for statecharts generate high-quality code from a model of a reactive system. However, they do not support runtime adaptation of the code because the code is generated monolithically at design time. The goal of this project is to develop an incremental code generator for statecharts that would support runtime adaptation. In other words, the behavior of a running system can be modified by changing its model on the fly.

Prerequisite knowledge: UML and OOAD

Difficulty Level: Medium to Difficult.