Interesting looking project from Matt Web. A desktop application that lets you "glance" at your friends.
This is the result of Matt asking: "What's the smallest scale of social interaction that can take place online?"
Interesting. I've been thinking along similar lines recently, wondering about other mechanisms for web annotation/path creation. Analysing little trails of browser activity across the web strikes me as very insect-like. Like a colony of ants interacting with it's environment. There must be a more mammalian metaphor, and the best I can think of is "scent marking".
The idea being that you can click a bookmarklet and record your scent ("I was here") against a page. Perhaps with some additional nuances such as your state at the time ("working", "playing").
Obviously scent marks aren't much use without a means to sniff the air. Checking whether a page is marked can be just another bookmarklet click. A more interesting angle is being able to just sample the air and decide which directions smell more interesting. I was thinking that a search engine interface is the closest to this, with a relevance ranking function based on the scents and whether you're part of the same group.
Posted by ldodds at September 15, 2003 09:44 AM | Feedback? | | TrackBackLeigh,
nice idea! I've been having some similar thoughts lately about FOAF (and collaborative filtering). But the problem I see (and that I think your scent trails might have) is that being part of the same group is not *necessarily* the same think as having similar interests.
Posted by: Steve Cayzer on September 16, 2003 08:02 AMI can see that this would break down where you're a member of multiple groups, as not all content would be relevant in all cases.
We can borrow a bit more from biology here and introduce the concept of group scents. If you've explicitly tied yourself to >1 foaf:Group then an additional option presented when recording a scent mark could control which group(s) scents are also recorded.
Everything still works in that case, and is another improvement over just path analysis as you're consciously recording a mark, rather than just having your footsteps recorded (which carry less information).
Posted by: Leigh Dodds on September 16, 2003 10:51 AMActually I'm not talking about multiple groups here. Even in a single group, saying "I belong to this group" is not necessarily the same thing as saying "I have the same interests/expertise/preferences as this group". It *may* be, depending on how you define the group. But not necessarily.
So what you're saying is that a scent mark is better than a footstep because
(a) It (may be) an explicit mark of approval (or disapproval). Stumble does this sort of thing.
(b) It is provenanced, which means you can say "Joe was here" or possible "A member of my foaf group was here" rather than just "somebody was here".
All of which I agree with. I'm just giving a mild warning that the provenancing provided by foaf groups may be misleading.
Cheers
Steve
Posted by: Steve Cayzer on September 16, 2003 11:10 AMIt's not quite a scent trail yet, but there're link loggers like http://del.icio.us that work on bookmarklets and categories. Would be nice to do a bit more with crossreferencing people with links and teasing out patterns
Posted by: l.m.orchard on October 7, 2003 03:47 AM