Just reading “Best software writing I” by Joel Spolsky.
Just about 2/3 of the way through and it talks of people interacting socially via the Internet. You look at forums, mailing lists, weblogs and sites like youtube and livejournal and you see they are all about the human interaction.
Places like friendsreunited.co.uk must have a nightmare time trying to get users to come back. Its not normal for you to think “I wonder where Kevin is now?” more than once a year and if you do find out thats it, curiosity killed. The site must have a limited appeal, it constantly needs new users.
Youtube is really just a beta website at the moment. Its small videos really make it hard to watch anything thats longer than a minute or two. The bandwidth and video phones just are not there yet. It has provided a good platform for proving what can be done as many videos have shown some of the sadder aspects of our humanity, things that would never make the 10 o’clock news.
The book also talks about reputations that underpin our interactions with each other. Linking that with an article about the Internet being 30 percent spam. It kind of shows that really what the web needs is a verification system that is dependant on reputation. We all know somebody and have a degree of trust with them. If I put up a website, my family could say they know me, my friends could as well. This could be done by simply linking to me (much as we do already). Businesses could allow employees to have a page on the company website where they could put links to themselves and other colleagues in the company and outside of it, thus building a map of trust.
Still with me?
Basically you are building up clusters of trusted people and businesses, the seven degrees of separation principal tells us we are all connected by seven other people. Spammers are easy to identify by human eyes and when they do it destroys not only the spam site but those that cluster to it. Now maybe you link to a spammer by accident? Simple just delink and you get some reputation back.
Reputation as within the brain isn’t a “yes, I trust this person” or a “no, I don’t”, its a grey scale. Also while you might trust your brother a lot someone in another country will only trust your brother a little due to a tenuous link. Everybody has their own measure of opinion of everyone else trustability (I think I just made a new word up).
Search engines clearly need to factor this in. Having just a vectored engine is great but us, the netizens, really need to think about how we interact with each other online. Removing the spam is a matter of trusting each other.
Trusting each other could be seen as a verification through time (as in “my trust is earned not given”). Do search engines include the time vector in their ranking? If www.hannent.eu has been on the net for about a week surely its not something that can be greatly trusted, however if it has been there a decade and has not been marked as spam then its clearly a good source and the links from it probably are as well.