dilton_dalton

 Profile Answers by dilton_dalton 

• Dec 8th, 2007

 

I am not sure why this question is under the topic of QA testing. 
Variance is the difference between what was planned and the actual value.  Two variance values related to the project tracking method known as "Earned Value" are the "cost variance" and the "schedule variance", both of which are expressed in monitary terms.  To understand what these terms mean, look up the earned value method in the internet.  (e.g. http://evm.nasa.gov/definition1a.html)  If effort variance is defined in a similar manner as the other two, then it would be the difference between the Effort actually expended to a given date less the Effort planned to be expended by the same date. 

Defect density is the number of defects divided by some measure of size.  The residual defect density is the number of defects remaining in a product at any stage if its development divided by the size.  There are a number of different ways of doing this calculation.  The defects may be any type of defect or they may be limited to functional defects (the kind that can be found by testing).  Size may be total size or it may be the sum of new and changed elements (e.g. lines of code or UML diagrams).  It is normally important that the size be trackable (e.g. lines of code produced can be tracked but number of function points produced cannot.)  There are several methods for estimating residual defects based on extrapolation of defect discovery rates.

Productivity is the ratio of output to input.  It can be expressed in any number of ways.  Revenue/employee is one that is used quite often at the corporate level.  Others used in software development include, Lines of code per developer month, Function Points per developer month, Test cases executed/tester month, ... The list of possible ratios is very large.  Generally these ratios are intended to be used for estimating purposes but many people are fearful that they will be used to compare the performance of individuals.  If that is done, then the people who are being measured will "game" the measurements so that they will look good which will often have a negative affect on the actual product.  As a general rule, metrics must never be used to reward or punish individuals.