This preview shows page 1. Sign up to view the full content.
Unformatted text preview: rements, and top level requirements, are initially the most interesting.
The process I use, is described in earlier INCOSE lectures [Gilb, SQC, 2005] and [Gilb 2005, CE
book, Spec QC chapter].
A small team of experts and managers will usually uncover 20-40 Major defects per page against
only 3 standards (clear, unambiguous, no-design in requirements). They are about 33% effective
in finding what is actually there at that moment. This can easily be proven by repeating the
review process, after the initial batch of defects is removed.
The simple conclusion, that my willing participant managers suddenly grasp, is that they have
been approving and accepting highly substandard requirements, and other specs, for years. They
are usually immediately ready to discuss how to rewrite the specs to a much cleaner level – like
maximum 1.0 majors per page remaining, statistically, and inferred.
9. Use metrics to prioritize, and determine priorities.
I argue that traditional weighting metrics are a very bad way of communicating priorities for...
View Full Document
- Spring '13
- Systems Engineering