2010 to 2019: Clouds, Crowds, Blogs, Big Data, and Predictive Analytics - The Technical and Social History of Software Engineering - page 329

4. Running meetings and other communications
5. Managing the project
6. Handling requirements creep
The SRM predictive analytics tool is aware of these historical software patterns
and therefore predicts (and measures) software defects found in requirements,
design, code, user documents, and bad fixes or secondary defects.
The SRM tool also predicts (and measures) the size and completeness of a vari-
ety of paper documents, including requirements, architecture, specifications, test
plans, status reports, user manuals, and many more. Paperwork costs are a major
factor in large-system development and especially so in defense software.
Requirements creep is a troubling phenomenon and averages perhaps 2% per
calendar month. In extreme cases, software projects have doubled in size after re-
quirements but before delivery. Therefore, dealing with requirements creep is a
necessary feature for software predictive analytics and is a standard feature of the
SRM tool.