290,000 file names. Each of the files on our laboratory's file server one particular afternoon appear in inode order (the order in which they were created) and a dot appears wherever two file names match. Unlike most of the other example dotplots shown here, this dotplot shows data associated with a program instead of a program's text (the file names are data associated with the file system program). The large white cross that spans the entire plot is caused by several thousand files with unique names that were created in succession. These file names were in fact generated by the UNIX split utility — someone was in the process of sending the X Window System to Spain via email! The checkerboard patterns along the main diagonal are names of multiple versions of files containing various large software systems. In this case, a dotplot of data associated with a program reveals two different ways that people use the program: 1) for temporary storage while transferring software, and 2) for maintaining multiple versions of software.