There is a theory(?) that "All the movie actors and actresses in the US constitute a 7-web."

If you start with actor A, who was in a movie with actress B, who was in a movie with actor C, etc, you can "link" any actor/actress who has ever been credited in a movie to any other actor/actress who has ever been credited in a movie - in no more than 7 links.

Some fanatics reportedly spend vast amounts of energy and effort trying to find "pairs" that require at least 8 links. Of course, when one of them reports an 8-link association, other fanatics must spend endless hours and energy trying to find a "shorter path" to connect that pair in no more than 7-links.

While the "shortest path" problem is usually considered as having no general solution other than by trial and error, statistical analyses have estimated a "web size" for the linked parts of the internet at something between a 17-web up to possibly a 23-web. (These numbers change frequently, depending on who you listen to.)

In principle, from any site (that has links to any other site(s)), you should be able to get to any destination site (that has links to other sites) in no more than 17 (or maybe it's 23) jumps.

This is, of course, only an "odds are" kind of estimate, since it's easy to build in "special sites" (with few links, usually) that are more remote.

A principal difficulty in making more specific estimates of the web-dimension comes from clusters of sites that have lots of internal links, but have few links to other clusters. Even a few "sparsely connected" clusters can greatly inflate the "shortest path" size of the web.

In additon, some estimates put the "webs" that have NO links to or from this "public internet" we live in at 500 to 800 times the size of "our internet." In many cases you can access this "other internet" if you know the address of a site there, but it's not generally accessible by "blue clicky."

John