Generering av API-dokumentation med hjälp av stora språkmodeller – Mot självförklarande API:er
Year: 2024
Type: dissertation
Abstract: Application Programming Interfaces (APIs) play a crucial role in modern software development. They are key in not only in enabling software reuse and adaptation, which greatly increase productivity, but also in facilitating exploration of new functionalities and technologies. The main way developers learn new APIs is through reading documentation. However, as many studies have shown, API documentation often suffers from a number of problems, such as incompleteness, outdatedness, or excessive verbosity. Consequently, efforts have been made to improve documentation and detect deficiencies, some leveraging Large Language Models (LLMs). This thesis explores the usage of LLMs to assist users learning new APIs. Employing Design Science, an artifact is developed, in the form of a web platform that leverages LLMs to generate API descriptions. The artifact is evaluated by a user study in the context of a university web development course. Among the 89 who students tested the platform, 58 provided feedback. The results show the potential of LLMs in generating and enhancing API documentation, thus facilitating API comprehension. A large majority of students find the LLM-powered additional support helpful with course assignments and useful for learning. However, students with greater programming experience tend to find less value in this supplementary resource, and prefer studying official API documentation instead.
Source: Aaltodoc (Aalto University)
Author: Jorelle, Yann
Institution: Aalto University
Language: English
Subfield: Information Systems
Field: Computer Science
Domain: Physical Sciences
SDG: Quality education
Open Access status: green
OpenAlex ID: W7056564542
