Scan represents Repository state at a given point of time. Scan contains one or more Source units. One Source unit can belong to multiple scans if it has not been changed between scans. Mesopotamia loads sources incrementally, which is important for codebases with thousands of source units (files).