Conclusion and outlook
In this article, the MAML framework was proposed to model mobile apps using a declarative graphical DSL. In contrast to the current practice of graphically configuring user interfaces by positioning UI elements on a screen-like canvas, MAML focuses on a process-centric definition of business apps. Using a sequence of platform-agnostic process elements, the notation aligns with the business perspective of managing processes and data flows, and makes app development accessible to domain users without software engineering experience. The approach is based on existing work on cross-platform business app generation and uses model-driven techniques to transform the models first into an intermediate textual representation before generating platform-specific source code. In particular, a data model inference mechanism was presented that enables real-time validation and consistency checks on partial data models, and overcomes the need for explicitly modeling a global data schema. Moreover, the inferred data can be used to provide contextual modeling support and enhanced semantic validation in the editor component. An empirical evaluation study supports the advantage of MAML over the related technical IFML notation, specifically with regard to its readability by domain experts. MAML therefore achieves the desired balance of abstracting programming-heavy tasks to understandable process flows while keeping the technical expressiveness required for automatic source code generation for multiple target platforms.
