The paper extends Reference Attribute Grammars with first-class relations (“Relational RAGs”), advancing a meta-language framework for defining incremental analyses over models/ASTs.
Its central scenario is continuous validation under ongoing model edits, addressing maintenance-like concerns through incremental recomputation strategies.
It reports empirical results (e.g., on benchmarked models/strategies) to quantify performance and responsiveness trade-offs for incremental validation.
