Evaluating a Graphical Model-Driven Approach to Codeless Business App Development

Abstract

Despite the growing interest in mobile app development, the creation of apps still follows traditional software development practices. Business apps are used by non-technical users in everyday work routines. However, their development is exclusively performed by software developers that need to centrally collect requirements and domain knowledge. Recent advances such as textual domain-specific languages (DSL) for cross-platform app generation reduce development efforts, but still focus on technical users. To alleviate these problems, the Münster App Modeling Language (MAML) is proposed as novel graphical DSL for specifying business apps. For each task to be accomplished within the app, the abstract process flows are modelled together with the respective data elements and view specifications in a combined model. Consequently, also non-technical users can express their domain knowledge without dealing with software engineering specifics. In contrast to existing process modelling notations, the MAML framework then allows for a codeless generation of apps for multiple platforms through model transformations and code generators. In order to automatically generate apps, the notation has to balance technical specificity and graphical simplicity. To assess the comprehensibility and usability of MAML's DSL, a qualitative usability evaluation was performed with software developers, process modellers, and domain experts.