Connect the dots puzzles with direction indicators

Abstract

Traditionally line puzzles or 'connect the dot' puzzles are solved by connecting numbered dots. These puzzles are solved by connecting points in a specific order indicated by numbers. These puzzles are handmade by placing dots on the outline of a drawing or line drawing. However there are some limitations and problems with this type of puzzle. For this thesis we researched an alternative puzzle type and propose an algorithm to extract a line puzzle from a given input, such as a line drawing. We researched the different aesthetics in this type of puzzle and a number of visual variations. Also we developed a proof of concept of an environment in which the algorithm can run. Using this proof of concept we processed a dataset with the algorithm and evaluate the resulting data and drawings. We conclude that there is a viable alternative for numbered 'connect the dot' puzzles that does not use numbered labels. For this alternative it is possible to generate puzzles by an algorithm that produces solvable and recognizable line puzzles with a running time of O(n² log³ n).