Bubble sort run against a shuffled list. Light and dark have been inverted from the original.
One of the traditional rites of passage that new computer programmers have to go through is learning a variety of functions called sorting algorithms. Arguably, sorting is a misnomer, as these functions deal with taking a list of items and putting them in the proper order, not separating them into different categories. They are used everywhere from ranking search results to alphabetizing address books.
There are dozens of these functions described in academia and in industry, all with subtly different properties. New sorting algorithms, and improvements to existing sorting algorithms, are an endless source of research papers. Some are faster than others, some require less memory than others, some are adapted to special hardware, and some are adapted to special datasets.
This is a project to create portraits of a variety of sorting algorithms. For each portrait, I started with one line of pixels, where each pixel is assigned a number. Larger numbers make brighter pixels, and smaller numbers make darker pixels. Then, I run the algorithm. Each time the algorithm moves items around in the list, I add another row to the picture showing the current state of the list. This continues until the list is entirely ordered, with black at the left, and white at the right.
I used several types of starting lists: I tried lists shuffled in random order. I tried lists set up in reverse order. I tried starting with an ordered list, and interleaving pixels from the two halves. I tried swapping the two halves. I tried swapping the first and last thirds. I even tried running the algorithms on an already sorted list--several algorithms insist on completely rearranging the list and then putting it back in order!
I also tried difference visualizations as well: For each pixel, I compared the current value of the pixel, and the value that it should have when the list is properly ordered. If they are the same, I color the pixel black. The more different they are, the brighter the pixel becomes. Areas that need work will light up, and areas that are already sorted go dark.
I've collected the most interesting pictures here. I have cropped or resized several, and I have tinted several with different colors that felt right. Enjoy.
Selection sort run against a list with swapped halves.
Insertion sort run against a list with the first and last thirds swapped.
Insertion sort run against a list with both halves swapped. Difference visualization.
Selection sort run against a shuffled list. Difference visualization.
