JDLX: visualization of dancing links

Data structures courses have settled on a familiar canon of structures and algorithms, and this is reflected in the standard textbooks. It is often useful for instructors to enliven such courses by presenting data structures that are of more recent interest, ones that may simultaneously challenge students' understanding of algorithms and their skills in programming. Exact cover problems, exemplified by the newly popular Sudoku game as well as the classic 8-queens problem, may be efficiently solved by the DLX algorithm popularized by Knuth in 2000, and this can provide a good capstone experience in a data structures course. The DLX algorithm operates by recursion on circular multiply linked lists. Because the pointer mechanics of the DLX algorithm is quite complicated, visualization techniques are called for. As the choreography of "dancing links" in DLX is highly visual anyway, this is very natural. In this paper we review best practices in algorithmic visualization for learners, and then describe a Java-based visualization of DLX applied to N-Queens. We also present some preliminary results that indicate that it is effective in enhancing student learning.