05 Problem Detection

Metrics and problem detection in software systems can help assess quality and facilitate improvement. While metrics provide measurable values, they have limitations including arbitrary thresholds and lack of context. Design problems like god classes and duplicated code negatively impact quality and maintenance. Detection strategies apply metric-based rules to find issues like high complexity, low cohesion, or excessive foreign data usage. Code duplication is analyzed at lexical, syntactical and semantic levels using techniques like string matching and tree comparison to identify significant duplicated code blocks. Metrics and detection strategies provide an initial understanding but their interpretation requires relating findings to design principles and domain knowledge.