How do you compute page rank for billion of pages that have cyclic links? Is that not the problem right after the initial idea?

▲

jmalicki 8 hours ago | parent [-]

That is not at all a problem. Links dampen their effect and you run it until convergence - read the paper. It's no different than summing an infinite convergent series, $\sum_{i=0}^n a^{-n}$.

Also, WTF, it seems impossible to find a PDF of the original paper still on the web without a paywall.

▲

hluska 7 hours ago | parent [-]

You’re underestimating the difficulty - actually downloading, parsing and updating the system with billions of pages was a very hard thing in the 1990s.

	▲	jmalicki 7 hours ago \| parent \| next [-]
		It's not an algorithmic issue, though, that's a systems issue. GGP was talking about the cyclic link nature which isn't an issue of size, it's an issue of convergence.
	▲	OroPla 5 hours ago \| parent \| prev \| next [-]
		In 1997, there were only a hundred thousand web pages. Not even millions. The one billion mark would not be breached until 2016 according to the results of a quick search.
	▲	icedchai 7 hours ago \| parent \| prev [-]
		Good thing there weren't billions of pages in the 90's.