Citation
Abstract
This article investigates the attributes of software project staffing and productivity implied by equating the effects of two popular software models in a small neighborhood of a given effort-duration point. The first model, the “communications overhead”’ model, presupposes that organizational productivity decreases as a function of the project staff size, due to interfacing and intercommunication. The second, the so-called “software equation,” relates the product size to effort and duration through a power-law tradeoff formula. The conclusions that may be reached by assuming that both of these describe project behavior, the former as a global phenomenon and the latter as a localized effect in a small neighborhood of a given effort-duration point, are that (1) there is a calculable maximum effective staff level, which, if exceeded, reduces the project production rate, (2) there is a calculable maximum extent to which effort and time may be traded effeetively, (3) it becomes ineffective in a practical sense to expend more than an additional 25-50% of resources in order to reduce delivery time, (4) the team production efficiency can be computed directly from the staff level, the slope of the intercommunication loss function, and the ratio of exponents in the software equation, (5) the ratio of staff size to maximum effective staff size is directly related to the ratio of the exponents in the software equation, and therefore to the rate at which effort and duration can be traded in the chosen neighborhood, and (6) the project intercommunication overhead can be determined from the staff level and software equation exponents, and vice versa. Several examples are given to illustrate and validate the results for use in DSN implementation.
Details
- Volume
- 42-72
- Published
- February 15, 1983
- Pages
- 70–77
- File Size
- 542.7 KB