Eliminating Setup Friction in PHP Web Development Labs Using Docker and DevContainers

<title>Abstract</title> <p>The first obstacle for students in a web development course is to set up a working PHP development environment. This has nothing to do with the real goal of the course: backend programming and client-server interaction. This paper describes a multi-semester case study with three cohorts and two universities in Bahrain. The case study compares a traditional XAMPP-based laboratory setup with an alternative based on Docker and DevContainers. The traditional setup had six constant friction points: differences in PHP versions, PDO/php.ini misconfigurations, inconsistencies in database credentials, IDE integration barriers, Mac vs. Windows inconsistencies, and a cyclical reinstallation spiral, which in some cases led to students losing a week’s worth of completed work. The problems related to the database were especially disruptive; there are twelve laboratory sessions in the course, eight of which require a functioning database connection to be established in week four, so one unresolved configuration error could potentially put two-thirds of the rest of the semester in jeopardy. The move from a traditional setup to a containerized environment decreased setup time from more than two hours to less than ten minutes, representing a reduction of over ninety percent. This decrease persisted regardless of whether the container environment was added at the beginning of the semester or in the middle of it as a corrective action. The container-based cohort of students carried out the laboratory activities in a timely manner and some of them self-reported that they did not need help anymore with their environment. This paper argues, based on cognitive load theory, that the advantages of containerization in education are better reflected through both the time efficiency and the lowering of extraneous cognitive load, which may threaten the entire semester on account of a single point of failure. Limitations and directions for future research are also discussed such as the need for a validated instrument to assess cognitive load.</p>