Expert judgement as an estimating method

This paper reviews some past research that may be relevant to the way that 'expert judgement' is used to produce estimates of software size and implementation effort. It then describes a survey of software estimators in one organization conducted via written questionnaires and some follow-up interviews. The survey paid particular attention to the information requirements of the estimators. A large proportion of the system development was found to involve the modification of existing systems and as a consequence a large part of the estimating process comprised the assessment of the impact of changes on existing code. The paper concludes by suggesting that some work could profitably be applied to attempting to support expert judgement rather than displacing it.