Event Details

Abstract:

Semantic nets are a useful modeling technique that can be used to representa wide va-riety of information. Rigi, SHriMP, and a number of other applications use this approach to abstract and visualize the structure of programs and other constructs. Thus far, most tools use variations on attributed, labeled, directed multigraphs to implement the seman-tic net. I will try to convince you that unattributed, unlabeled multi-hypergraphs provide a more uniform and powerful representation. I will explain my proposed model, demon-strate that it subsumes the old directed graph model, and show some of its advantages.

The talk will be accessible to everybody who has some experience with graphs and their use to model information. It will be focused on the abstract model, ignoring implementa-tion issues as much as possible. I also hope that the talk will be very interactive.