Domain invariant-based spreadsheet debugging

Spreadsheets are widely used in finance, accounting, auditing, etc. In our survey, we find that more than half of the spreadsheets used in daily audit work do not contain any formula. In practice, errors in such spreadsheets are usually handled manually according to the domain knowledge behind the data, which is time-consuming and easy to make mistakes. Inspired by this, we first propose to use domain invariants to do error detection and repair in spreadsheets. In this paper, a framework using domain invariants in spreadsheet debugging is proposed by incorporating with formal and prediction methods, respectively.