COINSE

CS453 Automated Software Testing, Spring 2023

Lectures

Time: 13:00-14:30, Mondays and Wednesdays Location: N1 102 (except on 28 February 2024, when it is in N1 117)

Lecturer

Shin Yoo shin.yoo@kaist.ac.kr Office: E3-1 Room 2405

Communication

All class announcements, as well as Q&A, will take place using a dedicated Slack workspace. You are required to join the class Slack workspace if you want to continue this course. It is strongly recommended that you install either a desktop or a mobile client, and get notifications. Email the lecturer or one of the TAs to get the invitation link, if you do not have one. When you sign up, please set your username as your real full name in English, followed by “(your student id number)”. For example, “Shin Yoo (20201234)”.”

Syllabus

This course is concerned with a broad range of software testing techniques, with a heavy emphasis on automation, tools, and frameworks, as well as the research outputs behind them. The topic will include, but are not limited to: black box testing/combinatorial testing, random testing, concepts of coverage, structural testing, mutation testing, regression testing, automated debugging, etc.

Prerequisite

  • Strong programming skills: you are required to actively contribute to group and individual project, which involves serious implementation. There will be also a number of hands-on sessions where we will program together during the class.
  • Unix/Linux-savvy: you should be familiar with the usual build tools and Unix/Linux command line environments.
  • Git-aware: knowing how to use git is mandatory for this course. First, we will use GitHub classroom for coursework. Second, you will be required to submit a github repository as part of your project deliverable.
  • Ideally, CS350 Introduction to Software Engineering.

Evaluation

Please note that, unlike previous years, we will have no exam. Instead, the course will have heavier emphasis on programming assignments. There will be also two quick quizzes held at random dates.

  • Coursework: 60%
  • Project: 30%
  • Quiz: 10%

References

We do not have a textbook per se, and the course will be based on slides and other reading material that are deemed appropriate. However, if you want to get broader sense for some of the topics dealt by this course, I recommend the following books and publications.

Lecture Schedule

Please note that the following schedule is tentative and may chance.

All assignments will be handled on GitHub Classroom.

Late Submission Policy

Submissions will be allowed up to a week after the initial deadline, but late submissions will be only given 50% of the grade they earn. All deadlines are announced at the beginning of the semester, so please make sure you allocate enough time for your assignments.

Assignment 0: GitHub Classroom Onboarding

You need to get familiar with GitHub Classroom: create a GitHub account if you do not have one, and learn the basics of Git. This assignment does not carry any grade, but unless you do this, we will not know your GitHub ID and therefore will not be able to grade your submissions. The assignment invitation link is here.

Assignment 1: Introduction to Metaprogramming

You will learn how to manipulate Python code using ast module. This assignment takes up 5% of total course grade. The aim of this assignment is to make you get familiar with AST-based program rewriting. The assignment invitation link is here.

Assignment 2: Python Coverage Profiler

Your task is to write a coverage profiler for Python that can measure statement and branch coverage: the goal is to replicate the same coverage as reported by coverage.py. This assignment takes up 20% of total course grade. The assignment invitation link is here.

Assignment 3: Concolic Engine

Your task will be to implement a lightweight concolic testing engine for a subset of Python syntax, using the peer architecture outlined in this technical report. This assignment takes up 15% of total course grade. The assignment link is here.

Assignment 4: Mutation Testing

Your task is to write a full mutation testing tool that mutates the give Python code, executes the given test cases against the generated mutants, and finally produces kill matrices. This assignment takes up 10% of total course grade. This assignment takes up 10% of total course grade. The assignment link is here.

Assignment 5: Delta Debugging

Your task will be to implement a delta debugging tool that minimises an error-revealing input. First, we will implement a linear and recursive DD for fake input. Subsequently, we will move onto Hierarchical Delta Debutting for Python programs (i.e., working with ASTs). This assignment takes up 10% of total course grade. The assignment link is here.

Project Aim

All teams should develop and/or implement an automated software testing technique based on an idea discussed during the course. I would encourage teams to pursue a novel idea, but a faithful reproduction of a state-of-the-art technique with solid evaluation would also do. If you are uncertain about your team’s idea, I will be happy to discuss it.

Proposal

All teams will give a presentation on 1st May to explain their project topics. I expect three things to be described clearly in the talk:

  • A testing problem the team aims to solve
  • The technique the team is proposing
  • A way of evaluation to show the proposed technique works and is competent

Team Project Deliverables

Everyone should submit the following, using the corresponding GitHub Classroom repo as the template:

  • the team report

  • the implementation: a public repository link in the report (e.g. GitHub or bitbucket repo) The team report should include:

  • a precise description of the problem you attempted to solve
  • a clear description of how you tried to solve the problem
  • a result of experimental comparison of before and after: in other words, what benefits did your solution bring?

Additionally, each individual member should submit a separate individual report in the repo:

  • details of what you have contributed to the project
  • peer assessment of your team members (yourself not included): use the scale of 10 to evaluate each of your teammates, and write clear justification for your score.

The submission deadline is the end of 13th June, UTC+9.

The final presentation dates for teams have been announced in the schedule section. Each team will have up to 15 minutes. If your team is scheduled on the early date, you can just report the progress up to that point, with a clear plan for the remaining work.

Teams

Form your teams by 24th April - write down the member names in the Google Sheet document (link will be available from the Slack workspace).

Examples from the previous years

I’ve picked a few projects from 2019 that I thought was interesting below.

Paper List