Results Balance between textual and visual notation Visual
Results • Balance between textual and visual notation – Visual notation typically works well at the macro level – Textual notation typically works well at the micro level – The user should have a choice between visual and graphical notation (similar to attributes vs. containment in UML class diagrams) – Different types of transformations may require different balance between visual and textual notations • Pattern of “model-to-model transformation(s) followed by a model-to-text transformation(s)” works. – The abstraction gap covered by model-to-text transformation(s) may vary depending on context Soft. Meta. Ware page 1
Results • Differentiation between model evolution transformations (same meta model for source and target) and language transformations (different meta model for source and target) with the latter case being the interesting one • Basic transformation constructs useful in practice are create, modify, delete. Modify is the complex case that requires attention. • Both imperative and declarative syntaxes need to co-exist • Research is required to determine heuristics for classifying transformation syntaxes as “good” or “bad”. • Textual languages (programming languages, template languages) work well in the small. Graphical languages that raise the level of abstractions improve scalability. Soft. Meta. Ware page 2
Results • We need escapes/hooks in transformation languages that allow to switch between graphical and textual syntaxes, and even to OO programming languages. • We need to collect benchmark examples of practically relevant transformations to discover paradigms for transformation management Soft. Meta. Ware page 3
- Slides: 3