[Date Prev] [Date Next] [Thread Prev] [Thread Next] Indexes: Main | Date | Thread | Author

Re: [ba-unrev-talk] More on backlinks


At 11:04 AM 5/17/2002 -0700, you wrote:
>http://www.kuro5hin.org/?op=displaystory;sid=2002/5/16/152614/482
>"The idea for this article came about the other day while I was viewing 
>some of the recent posts on drop.org. I followed a link to an article, 
>'Blogspace Under the Microscope', on the O'Reilly Network. The author was 
>talking about a 'mutation' and the next 'phenomenon' in weblogging. He was 
>referring to 'backlinks' - links which point back to the referring web 
>page - which the web site Disenchanted has been using for about a year. 
>The article consisted of very few links and no historical or background 
>information to the idea of 'backlinks' and implies that Disenchanted 
>invented this idea; which is dismissed by an accompanying article on 
>Disenchanted. "
>
>More at http://www.foresight.org/WebEnhance/backlinks.news.html
>which is part of the http://www.foresight.org/WebEnhance/enhance.html 
>WebEnhance project.  Pretty old stuff there.    (01)

It should be noted that these folks http://pespmc1.vub.ac.be/LEARNWEB.html 
have been advocating using Hebbian learning with links as a means of 
"learning" what's relevant.  Hebbian learning is a mechanism 
http://www.nd.com/definitions/hebbian.htm that tweaks "weights" on links: 
the more a link is active, the stronger its weight becomes.  That idea has 
been modified over time to "competitive" and "differential" 
versions.  Differential applies a "time" window such that, if someone uses 
Link A to reach B, and someone else uses Link A to reach B, and that 
happens within the window, then weight of Link A is enhance, otherwise 
nothing happens to the weight.    (02)

Imagine what would happen during a "slashdot" event!  You might never get 
rid of that link!    (03)