Making Systems Explainable
Date:
This was an invited Keynote presentation at VISSOFT 2022 (10th IEEE Working Conference on Software Visualization), Limmasol, Cyprus.
Abstract
What makes software systems explainable? As we develop and maintain software, we have questions to ask about the code, but piecing together the answers remains hard. The main interface the classical IDE offers is a text editor for the source code. Code, documentation, and the running system are disconnected. In this keynote presentation, we will show how software systems can be made explainable with the help of three interacting technologies: (i) live notebooks that can be used to create narratives that link documentation, source code, and running applications, (ii) example methods that not only perform tests, but produce live examples that can be used within narratives, to explain use cases, scenarios and features, and (iii) a moldable inspector that can be easily extended with live custom views to answer domain-specific questions about software systems. With the help of running examples we will show how these technologies work together to provide a radically different kind of development experience.