摘要: Abstract Since the first investigations on web-graph compression, it has been clear that ordering of nodes a web graph fundamental influence compression rate (usually expressed as number bits per link). The authors LINK database [Randall et al. 02], for instance, investigated three different approaches: an extrinsic (URL ordering) and two intrinsic orderings based rows adjacency matrix (lexicographic Gray code); they concluded URL many advantages in spite small penalty compression. In this paper we approach issue more systematic way, testing some known proposing new ones. Our experiments are made WebGraph framework [Boldi Vigna 04], show technique structure can produce significantly results. particular, transposed graph, is less effective, mixed orderi...