Master's Thesis of Lucas Köhler

A Model-Driven JSON Editor

Many engineering domains require the input and modification of structured data (also known as models). This structured data is usually defined by a meta-model (e.g. in JSON Schema). To modify instances of structured data, users require proper tooling. The manual implementation of suitable tooling is costly, leads to re-implementing common features, and causes additional development effort whenever the meta-model changes.

In this thesis, we develop a model-driven framework to semi-automatically create an editor for a given meta-model. This editor allows the creation and modification of structured data as specified by the meta-model. The goal is to minimize the manual work required to create and maintain such an editor.

As a preparation, we conduct an extensive analysis of the requirements of structured data tools. First, we conduct a literature review on multilevel modeling tool requirements. Second, we complement the results with requirements gathered by a tool analysis of nine data editors. Based on this, we design and implement the model-driven editor framework. Thereby, we focus on allowing to create an editor with minimal effort while simultaneously providing extensive options to customize and extend the generated editors.

We evaluate our framework by configuring and creating three editors for the existing meta-models Ecore, JSON Schema, and UI Schema. Finally, we successfully valuate the usability of editors created by our framework by conducting a System Usability Scale test for the UI Schema Editor. It achieves a score of 79.5 out of 100 points.