Cluster-based find and replace

Abstract

Current Find and Replace tools offer their users only two options when performing a find and replace: replace one match at a time while prompting the user to confirm each replace, or replace all matches at once without any user confirmation. Each of these options can be prone to errors. In this thesis, I investigated, implemented, and tested a third option when performing a find and replace: a cluster-based find and replace. Instead of replacing one match at a time or all matches at once, the user can choose to replace a cluster of similar matches. While I hypothesized that a cluster-based find and replace interface would make users complete a find and replace task faster and more accurately, preliminary user studies suggest that a cluster-based interface may improve speed.