Indexing the archive…
Your Universe of Digital Possibilities
A blind surfer wanders a web of links: most of the time it follows a random link, and now and then it teleports to a random page. Let it run forever and the fraction of time spent at each node settles into a single number — rank. It is importance defined in a loop: you matter if pages that matter point at you. That fixed point is the dominant eigenvector of the Google matrix, and just multiplying by the matrix again and again finds it. The same idea reads signal out of any web of relationships — citations, roads, money, a knowledge graph — which is exactly what a Lattice of data, models and tools is.
A node’s rank is the rank poured into it along incoming links, each source splitting its own rank evenly across its out-links. Importance is recursive — you matter if things that matter point at you — plus a small flat share (1−d)/N so nothing starves.
M is the link matrix (a column per node, its 1s split evenly down its out-links). The teleport term — jump to a random node with probability 1−d — makes G positive and connected, so a single, well-defined rank is guaranteed to exist.
Just apply G again and again: any starting guess is dragged onto G’s leading eigenvector. It converges at the spectral gap — about a factor d per sweep — which is why ~50 passes rank the whole web.
The rank is the equilibrium of a Markov chain — the fraction of eternity a blind random surfer spends at each node. Run one walker forever, or solve for the fixed point: the same vector. Kin to The Walk, set on a graph.
A positive matrix has one largest eigenvalue, and its eigenvector is all-positive and unique. That theorem is the guarantee under PageRank: exactly one rank vector, every entry a real fraction, and power iteration always finds it.
This is the rack’s network instrument — structure read straight off a web of relationships. The surfer is one of The Walk’s (INST·19) random walkers, set loose on a directed graph instead of a line, and the rank is its stationary distribution. It shares its substrate with The Contagion (INST·28), which runs a different process — spreading, not ranking — over the same kind of graph, and its emergence-of-one-number with The Chorus. Most of all it is the graph twin of The Shadow (INST·31): one recovers the hidden structure of a time-series, the other the hidden structure of a network. Together they are the Perception Engine’s two halves — signal and relation — and the maths under a Lattice of data, models and knowledge.