Improved feedback and debugging support for student assembly programming

Liu, Tingkai

Permalink

https://hdl.handle.net/2142/110283 Copy

Description

Title

Improved feedback and debugging support for student assembly programming

Author(s)

Liu, Tingkai

Contributor(s)

Lumetta, Steven

Issue Date

2021-05

Keyword(s)

• Symbolic execution
• Reverse execution
• Student feedback
• Debugging

Abstract

Debugging is one of the most difficult tasks in programming, and students in programming classes often struggle with this process. We have developed advanced debugging aids for our students, such as language-specific static analysis embedded into editors and the use of symbolic testing to identify behavioral differences between student code and a correct solution. But even with these advances, the debugging process is still challenging for students. In this thesis, we introduce two improvements to the debugging process. First, we improve the feedback provided from symbolic execution of student code by leveraging the availability of information about code structure and functionality within the symbolic testing engine. Specifically, we analyze the relationships between input subspaces for which the student code behaves correctly and subspaces for which it does not, then use the relationships between those subspaces to report errors to the students. Second, we add support for reverse execution to the debugger that students use to test and debug their code. These improvements are implemented and tested in the the tools that we use to teach LC-3 assembly language programming, but the ideas can also be applied to other languages. We illustrate the value of leveraging code structure using samples of student code submitted for the LC-3 programming assignments in Fall 2020 ECE220 at ZJU-UIUC Institute, for which students individually wrote over 690 (median value) lines of LC-3 assembly code. Students in the class made use of reverse debugging when working on their final LC-3 assignment and rated it highly in an anonymous survey.

Type of Resource

text

Language

en

Permalink

http://hdl.handle.net/2142/110283

Owning Collections