Smooth Linear Approximation of Non-overlap Constraints

  • Authors:
  • Graeme Gange;Kim Marriott;Peter J. Stuckey

  • Affiliations:
  • Department of Comp Sci and Soft Eng, University of Melbourne, Australia 3010;Clayton School of IT Monash University, Australia 3800;Department of Comp Sci and Soft Eng, University of Melbourne, Australia 3010 and National ICT Australia, Victoria Laboratory,

  • Venue:
  • Diagrams '08 Proceedings of the 5th international conference on Diagrammatic Representation and Inference
  • Year:
  • 2008

Quantified Score

Hi-index 0.00

Visualization

Abstract

Constraint-based placement tools and their use in diagramming tools has been investigated for decades. One of the most important and natural placement constraints in diagrams is that their graphic elements do not overlap. However, non-overlap of objects, especially non-convex objects, is difficult to solve and, in particular, to solve sufficiently rapidly for direct manipulation. Here we present the first practical approach for solving non-overlap of possibly non-convex objects in conjunction with other placement constraints such as alignment and distribution. Our methods are based on approximating the non-overlap constraint by a smoothly changing linear approximation. We have found that this in combination with techniques for lazy addition of constraints, is rapid enough to support direct manipulation in reasonably sized diagrams.