The paper describes a framework to run experiments regarding flaky tests for Python projects. Three tools were built for that purpose: the framework, ShowFlakes, and TestInspect.

Sixteen features were used to classify OD and NOD test cases.

Test rerun is not enough to detect flaky tests? For OD, indeed, plain reruns maybe not be enough, as they will not pollute the running context. However, if the running configuration is stable, is that a problem?

Update my personal library of papers on flaky tests (backward snowballing).

Read “Teaching data representation using real-world applications” and used it as inspiration to create the concept map on data representation. It was a nice touch that the paper used bit flippers (“Flippy Do“, from Code.org) to teach binary encoding of decimal and fractional numbers, and the Mars 2020 Lander parachute to teach encoding of textual data.

While reading the previous paper, I stumbled on some papers about propositional logic: “InfoTraffic: teaching important concepts of computer science and math through real-world examples” (10.1145/1227504.1227349). It provides a nice online application to learn about propositional logic, using road traffic control as metaphor: https://logictraffic.ch/.

Read “An online tool for easy-to-set-up, visualizer-based, and auto-gradable full tracing exercises”. It is well suited to program understanding and debugging competences, as required in the Programming Fundamentals exam I applied to my students last Monday. The tool. TracingQuiz, builds upon PythonTutor's visualizer, requiring the student to answer a quiz for each line under execution. It should be available at https://tracingquiz.xyz/, but I couldn't access it (access denied).