Apparatus, processes, and related technologies for comparison between a target item of software code and a reference set of software code. The target item is preprocessed to be compared against a reference item from the reference set to identify a selected set of lines of software code from the target item to be used for the comparison. Each line of the selected set of lines from the target software item is individually compared with lines of software code from the reference set to produce a measure of similarity between the target software item and at least one reference item of software code from the reference set. Various techniques for maintaining and updating a numerical representation of similarity of the target item with each reference item, the numerical representation being stored in a corresponding element of a data structure.