- From: Gordon Joly <gordo@dircon.co.uk>
- Date: Wed, 8 Nov 2000 09:30:38 +0000 (GMT)
- To: Matt Jensen <mattj@newsblip.com>
- cc: William Loughborough <love26@gorge.net>, Craig Pugsley <craig.pugsley@mimesweeper.com>, "'www-rdf-interest@w3.org'" <www-rdf-interest@w3.org>, "'semantic-web@w3.org'" <semantic-web@w3.org>
Jakob Nielsen shows that both "page requests" (for a site such as http://www.sun.com/) follow the Zipf law. He states (on the site you gave, http://www.useit.com/alertbox/zipf.html) "The figure shows incoming page-requests to a single site. Other studies have found that Zipf curves characterize the outgoing page requests from the employees of an organization (there are a few pages that everybody look at and a large number of pages that are seen only once)." So the law applies for "page popularity" plotted against "hits per month" (his terminology). *****>>>>> earlier it was said... >However, because the Web is described by Zipf's law [1], you might get >good value out of convincing "only" the top 100 sites (in page views) to >implement your semantic system. Then a form of Metcalfe's law [2] >applies, where other web sites have more and more reasons to join. Yes, to convince them. But like HTML (the well know computer virus), inertia rules. Did PNG take off like a rocket? Answer - no. And that was a very simple in comparison. Gordo. On Tue, 7 Nov 2000, Matt Jensen wrote: > No, I haven't proved it, but others have, with regard to requested URLs. > > Scroll down in footnote [1] until you see the section titled "Zipf's law > in Web Access Statistics." There are 18 academic papers cited. Probably > the most popular non-academic page describing this is Jakob Nielsen's > page, "Zipf Curves and Website Popularity", at > http://www.useit.com/alertbox/zipf.html. > > -Matt Jensen > NewsBlip > Seattle > > > On Tue, 7 Nov 2000, Gordon Joly wrote: > > > > [1] http://linkage.rockefeller.edu/wli/zipf/ > > > > Have you proved that the described by Zipf's law? With which > > parameters? Links? Pages? Pageviews? > > > > Gordo. > > > > -- > > Gordon Joly http://www.pobox.com/~gordo/ > > gordo@dircon.co.uk gordon.joly@pobox.com > > > > -- Gordon Joly http://www.pobox.com/~gordo/ gordo@dircon.co.uk gordon.joly@pobox.com
Received on Wednesday, 8 November 2000 04:54:19 UTC