varying definitions of " scale free just more heavy-tailed than exponential, Pr(k)xe-Ak not Pr(ki) o k-a exactly, but more plausible than alternatives lausibly Pr(k)ok a for any a · plausibly Pr(k)∝k- but for2<a<3only · plausibly Pr(k)∝ k but only in upper tail plausibly Pr(k)∝k-a except for exponential cutoff scale free is well defined mainly for simple graphs