Exploring conflict reasons for graph transformation systems

Paper i proceeding, 2019

Conflict and dependency analysis (CDA) is a static analysis for the detection of conflicting and dependent rule applications in a graph transformation system. Recently, granularity levels for conflicts and dependencies have been investigated focussing on delete-use conflicts and produce-use dependencies. A central notion for granularity considerations are (minimal) conflict and dependency reasons. For a rule pair, where the second rule is non-deleting, it is well-understood based on corresponding constructive characterizations how to efficiently compute (minimal) conflict and dependency reasons. We further explore the notion of (minimal) conflict reason for the general case where the second rule of a rule pair may be deleting as well. We present new constructive characterizations of (minimal) conflict reasons distinguishing delete-read from delete-delete reasons. Based on these constructive characterizations we propose a procedure for computing (minimal) conflict reasons and we show that it is sound and complete.