Applied Software Measurement by Capers Jones

A history of software development management from the numbers perspective. Basically, until the early 80's, we didn't use semantic analysis to gauge productivity.

It appears that, for knowledge workers such as software professionals, the impact of physical office environments or productivity may be as great as the impact of tools and methods.

-- Introduction

PeopleWare strikes again

function points / staff-month average in 1990 = 5, 82% of all projects >= 2, 18% of all projects >= 16

-- Introduction, Chart of Function Points / Staff-Month in the US

IBM was also the first company to discover that software quality and software productivity were directly coupled and that the projects with the lowest defect counts by customers were those with the shortest schedules and the highest development productivity rates.

-- Introduction

umm. complexity kills?

The companies that wish to improve but do not measure are at the mercy of fads and change. Progress may not be impossible, but it is certainly unlikely. Only when software engineering is placed on a base of firm metrical information can it take its place as a true engineering discipline rather than an artistic activity, as it has been for much of its history.

-- Introduction

DeMarco's Bang Metric was the first attempt to apply functional metrics to the domain of systems and scientific software. Metrics = Create, Replace, Update, Delete on
functional primitives, data elements, objects, relationships, finite states and their transitions.

-- History and Evaluation of Functional Metrics

English has 51 statements per function point

-- History and Evaluation of Functional Metrics, Table of Statements / Function Point for some Languages

Most forms of testing are less than 30% efficient, in that they will find less than 1 bug in 3 present. Formal design and code inspection tend to be most efficient, and they alone can exceed 60% in defect removal efficiency.

US Averages for Software Productivity and Quality