Date of Award
Summer 8-15-2018
Degree Name
Doctor of Philosophy (PhD)
Degree Type
Dissertation
Abstract
Programmers in artifact-based contexts could likely benefit from skills that they do not realize exist. We define artifact-based contexts as contexts where programmers have a goal project, like an application or game, which they must figure out how to accomplish and can change along the way. Artifact-based contexts do not have quantifiable goal states, like the solution to a puzzle or the resolution of a bug in task-based contexts. Currently, programmers in artifact-based contexts have to seek out information, but may be unaware of useful information or choose not to seek out new skills. This is especially problematic for young novice programmers in blocks programming environments. Blocks programming environments often lack even minimal in-context support, such as auto-complete or in-context documentation. Novices programming independently in these blocks-based programming environments often plateau in the programming skills and API methods they use. This work aims to encourage novices in artifact-based programming contexts to explore new API methods and skills. One way to support novices may be with examples, as examples are effective for learning and highly available. In order to better understand how to use examples for supporting novice programmers, I first ran two studies exploring novices' use and focus on example code. I used those results to design a system called the Example Guru. The Example Guru suggests example snippets to novice programmers that contain previously unused API methods or code concepts. Finally, I present an approach for semi-automatically generating content for this type of suggestion system. This approach reduces the amount of expert effort required to create suggestions. This work contains three contributions: 1) a better understanding of difficulties novices have using example code, 2) a system that encourages exploration and use of new programming skills, and 3) an approach for generating content for a suggestion system with less expert effort.
Language
English (en)
Chair
Caitlin Kelleher
Committee Members
Roger Chamberlain, Sanmay Das, Bjoern Hartmann, Alvitta Ottley,
Comments
Permanent URL: https://doi.org/10.7936/ee9y-sm80