Lines of code can be useful as a software size, productivity, estimation, and defect-density metric only when the counting method is consistent. Software quality engineering requires predefined counting criteria because teams may otherwise count physical lines, logical statements, comments, blank lines, generated code, reused code, or executable lines differently. Without predefined rules, comparisons across modules, projects, teams, or releases become unreliable. Requirements do not normally define how code lines should be counted. A line-by-line code review may inspect code quality but does not establish a measurement standard. Module interfaces describe interactions, not counting rules. Therefore, lines-of-code metrics must be based on predefined criteria so the data is repeatable, objective, and suitable for analysis.

================