A motor for our kayak: Aggregated Trust Networks

The ebiquity blog at UMBC wonders if it possible to get A motor for our kayak, based on the metaphor that Clay Shirky raised, ( and further elaborated on by Cory : Net is a kayak, driven by its environment .)

Given that an increasing population of url’s and users at (to take one example) del.icio.us, it usefulness as a discovery engine might/will decline (lower quality material, more tag spam), how to counter that trend? One (obvious?) way is to use a “trust network”: people whom opinions we trust, like friends and families, but also people whom we have discovered as having judgment (or taste) we agree with. Of course this would lead to a emergence of the Power Law in blogs, links and Tags, as feared by some.

The ebiquity blog suggests using “RDF information on our social network (FOAF, (Friend Of A Friend))”, but also cautions: “Personalized and topic-based ranking offers many challenges” and suggests a another resource: An Analytical Comparison of Approaches to Personalizing PageRank (PDF).

A fancy automated system for the (here come the buzz words) Aggregated Tags of my Trust Network, with weighing, is many days off. (Or a least a couple, given the speed of innovation on the net these days). However even a simple form of Trusted Tag discovery is possible given that del.icio.us allow of RSS feeds of another users bookmarks. So for now, I have in my aggregator the feeds for Jon Udell, Jeffrey Veen and others (plus my own: FalsePositives). (the Feed url is at the bottom of each page).

Might it soon be possible to Aggregate my Blogroll (via a linked OPML file, which is a sort of FOAF) and aggregate their Blogroll’s, Links and entries. You would what to set a limit on how deep to go (given the Six Degrees of Connection phenomena). It would nice to use duplications to add weight as well as the Vote=”+” attribute and the new rel=”NoFollow” attribute. Altogether maybe we go all the way and use whuffie=”10″ attribute, as I found suggested on Jo Ito’s Blog :adding more information to links. (I was going there away, implicitly) Finally, We could use the rel=”tag” attribute to discover how a link or blog entry has been Tagged and Classified, but not in a manner explicitly linked to technorati.

All this aggregated, page ranked, and Tag mapped, as part of my own personalized Google Desktop, or FireFox extentsion? All the pieces are there, if not widely implemented. I shall leave the actual construct, which is merely a matter of programming, as a exercise to the reader ;).

categories:folksonomies / / Semantic Web / Firefox / Google

Leave a Reply