
84 ◾ Web 2.0 and Beyond: Principles and Technologies
skewed distribution form (Clauset et al. 2009). ere are oen debates within scientic
disciplines as to which is the best t for a particular set of data and Mitzenmacher (2001)
for example, notes the diculty of deciding whether income distribution is a log-normal or
a power law. As Newman (2005, 8) writes: “the scientist confronted with a new set of data
having a broad dynamic range and a highly skewed distribution should certainly bear in
mind that a power-law model is only one of several possibilities for tting it”.
7.2.2 Scale-Free Networks
Let us return to the work of Barabá