A review of studies on expert estimation of software development effort

摘要: This paper provides an extensive review of studies related to expert estimation software development effort. The main goal and contribution the is support research on estimation, e.g., ease other researcher's search for relevant studies. In addition, we provide practitioners with useful guidelines, based research-based knowledge processes. results suggest that most frequently applied strategy projects, there no substantial evidence in favour use models, are situations where can expect estimates be more accurate than formal models. following 12 ''best practice'' guidelines evaluated through review: (1) evaluate accuracy, but avoid high evaluation pressure; (2) conflicting goals; (3) ask estimators justify criticize their estimates; (4) irrelevant unreliable information; (5) documented data from previous tasks; (6) find experts domain background good records; (7) Estimate top-down bottom-up, independently each other; (8) checklists; (9) combine different strategies; (10) assess uncertainty estimate; (11) feedback accuracy task relations; and, (12) training opportunities. We found supporting all principles, suggestions how implement them organizations.