Live Model Transformations Driven by Incremental Pattern Matching

In the current paper, we introduce a live model transformation framework, which continuously maintains a transformation context such that model changes to source inputs can be readily identified, and their effects can be incrementally propagated. Our framework builds upon an incremental pattern matcher engine, which keeps track of matches of complex contextual constraints captured in the form of graph patterns. As a result, complex model changes can be treated as elementary change events. Reactions to the changes of match sets are specified by graph transformation rules with a novel transactional execution semantics incorporating both pseudo-parallel and serializable behaviour.