A Preliminary Inquiry Into the Software Estimation Process
The report reviews the literature of software estimation and reports a small experiment comparing Delphi with face-to-face group judgment to predict the time necessary to program an information system--in this case, the Air Force's PDSO (Personnel Data System--Officers). Planning software production is necessary but almost impossible at present. Cost to completion depends on many factors, some unknown at the time and all hard to quantify (the difficulty of the task, the programmer's ability and familiarity with the procedures involved, the degree of definition provided him, and about 80 other factors). The experiments undertaken failed to establish the utility of either estimation method. Primary recommendations for further research are: (1) more effective data collection, (2) analysis of characteristics of good estimators, and (3) formal inquiry into the techniques used by estimators. (Author).
What people are saying - Write a review
We haven't found any reviews in the usual places.