Abstract In Model-Driven Engineering, editing models is typically not merely a purely textual endeavor, but rather a mix between textual and graphical editors and views. Both have their advantages and use cases where either textual or diagrammatic representations are better suited to edit and understand models. Therefore, a modeling framework offering the best of both worlds can be advantageous. We define the text-first approach to combine the textual and diagrammatic representations by automatically synthesizing the textual model into a diagram. We present the KIELER text-first diagramming framework and its take on current challenges for model visualization and compare it to the diagram-first approach, as exemplified by the GLSP framework.

Abstract In 2010, Fuhrmann et al. argued for enhancing modeler productivity by providing tooling that, put simply, combines the best of textual and graphical worlds. They referred to this as pragmatics , and argued that a key enabler would be the ability to automatically synthesize customized graphical views from a (possibly textual) model. The model would be the “ground truth” used, for example, for downstream code synthesis and simulation; the graphical views would typically be abstractions from the model serving various purposes, including documentation. Twelve years later, we reflect on their proposal, and illustrate the current state with the recently developed polyglot coordination language Lingua Franca (LF). LF has been designed with pragmatics in mind since early on, and some characteristics of LF make it particularly suited for pragmatics-aware programming and modeling. However, the underlying pragmatic principles are broadly applicable, and by now a set of mature open source tools is available for putting them into practice.