
Analyzing Software Measurement Data ◾ 245
four parts. We dene the box length, d, to be the distance from the upper
quartile to the lower; thus, d = u − l. Next, we dene the tails of the dis-
tribution. ese points represent the theoretical bounds between which
we are likely to nd all the data points if the distribution is normal. If
the data is on an interval, ratio, or absolute scale, the theoretical upper
tail value is the point u + 1.5d, and the lower tail value is l − 1.5d. ese
theoretical values must then be truncated to the nearest actual data point
to avoid meaningless concepts (such as negative LOC) and to demonstrate
the essentia ...