Software measurement, like measurement in any other discipline, must a
dhere to the science of measurements if it is to gain widespread accep
tance and validity. The observation of some very simple, but fundament
al, principles of measurement can have an extremely beneficial effect
on the subject. Measurement theory is used to highlight both weaknesse
s and strengths of software metrics work, including work on metrics va
lidation. We identify a problem with the well-known Weyuker properties
, but also show that a criticism of these properties by Cherniavsky an
d Smith is invalid. We show that the search for general software compl
exity measures is doomed to failure. However, the theory does help us
to define and validate measures of specific complexity attributes. Abo
ve all, we are able to view software measurement in a very wide perspe
ctive, rationalising and relating its many diverse activities.