Project Measurements

Successful large projects are most often found in companies that have software measurement programs for capturing productivity and quality historical data , Thus any new project can be compared against similar projects to judge the validity of schedules, costs, quality, and other important factors. The most useful measurements for projects in the 10,000-
function point domain include measures of the following:

• Accumulated effort.
• Accumulated costs.
• Development productivity.
• Volume and rate of requirements changes.
• Defects by origin.
• Defect removal efficiency.

Measures of effort should be granular enough to support work breakdown structures. Cost measures should be complete and include development costs, contract costs, and costs associated with purchasingn or leasing packages. There is one area of ambiguity even for top companies
and successful projects: The overhead or burden rates established by companies vary widely. These variances can distort comparisons between companies, industries, and countries, and make benchmarking difficult. Of course, within a single company this is not an issue.

The federal government, some military projects, and the defense industry still perform measurements using the older lines-of-code metric. This metric is hazardous because it cannot be used for measuring many important activities such as requirements, design, documentation,
project management, quality assurance, and the like. There are also programming languages such as Visual Basic that have no effective rules for counting lines of code. About one third of the large software projects examined utilized several programming languages concurrently, and one large application included 12 different programming languages.

Measures of quality are powerful indicators of top-ranked software producers and are almost universal on successful projects. Projects that are likely to fail, or have failed, almost never measure quality. Quality measures include defect volumes by origin (i.e., requirements, design, code,bad fixes) and severity level, defect severity levels, and defect repair rates. really sophisticated companies and projects also measure defect removal efficiency. This requires accumulating all defects found during development and also after release to customers for a predetermined time period. For example, if a project finds 900 defects during development
and the users find 100 defects in the first three months of use, then it can be stated that the project achieved a 90 percent defect removal efficiency level. Of course, any defect found after the first three months lowers the defect removal value.