Visualizing Answer Set Programming.
Arne König · Theory and Practice of Logic Programming · 2013
A distinguishing feature of Answer Set Programming (ASP) is its declarativity, decoupling problem representations from problem solving algorithms. However, this strict separation should not lead to viewing the solving process as a magic black box refusing any insights into how the problem is solved. My thesis project aims at creating visualization tools suitable to enhance the transparency of ASP systems. The first part presented here focuses on the solving of propositional ASP programs and consists of a flexible data logger protocoling relevant events occurring during ASP solving as well as a visualization back-end offering various views on both the underlying problem structure and the solving process. The presented system allows to re-connect the solving process with the original problem specification and thus to reveal how the original problem is actually solved. However, it is currently aimed at ASP developers and contains no information from the grounding process that transforms the first-order specification into the propositional encoding given to the solver. My current research focus is on incorporating information from the grounding phase to make the tools usable for users who usually do not necessarily have the same level of experience with ASP.