February 4, 2004
Social Networking
So apparently Google launched their version of Friendster, called Orkut. You need an invitation to join, so I haven't gotten to use that system, but I did sign-up at Tribes and Friendster.
This whole social networking thing is very very interesting but I haven't really seen it work very well yet. The main problem is that social networks are inherently implicit; I don't go around saying to people "you're a good friend," or "you're just an acquaintance;" these things are implied.
The problem with sites like Friendster/Tribes is that there is no way to distinguish the various people in your life; someone is your friend or they're not.
This causes a problem; I have people I would consider "real" friends while others are mere locational friends (friends because they attend UNC, etc.). If I wanted to really analyze my social network, I would be able to distinguish the two. Don Park gets it somewhat close with the friendship circles, but that's still explicitly defining your relationships (albeit if you built a desktop app you could do drag and drop with thumbnails).
What I want to see is some sort of superproject (by Ph.Ds) that implements FOAF and uses some sort of program like Dashboard that constantly trawls your workspace and stores data.
For example, you could easily use the FOAF files to associate IMs/e-mails with existing users; if the program could also remember when these interactions occur, it could also remember the frequency of the contacts.
For example, if user "ABAlex " messages me in AOL Instant Messnger, a Dashboard-like program would pick up on this, grab my FOAF file and do a quick search of my friends to see if it can pick up on who ABAlex is. It would quickly find out that it's Borst ... and then store some data saying "Hey, Roy chatted with Borst on this date..." Everytime someone new messaged me or e-mailed me, the program would pick up on this and then implicitly remember that I am a "friend." Perhaps the engine might start using a simple frequency analyzer (# of interactions over a time span) to say that this person would be a "closer associate."
Of course, the downfalls of this? It's only limited to computer interaction; a lot of my 'real' friends don't bother with computers ... and how do you factor in phone conversations? Etc. etc.
Ok I'm not really making sense here. I have an exam in 5 hours and I'm up rambling about stuff like this again. Gotta go study ... gonna post this entry to the future. Cya in the future!
This whole social networking thing is very very interesting but I haven't really seen it work very well yet. The main problem is that social networks are inherently implicit; I don't go around saying to people "you're a good friend," or "you're just an acquaintance;" these things are implied.
The problem with sites like Friendster/Tribes is that there is no way to distinguish the various people in your life; someone is your friend or they're not.
This causes a problem; I have people I would consider "real" friends while others are mere locational friends (friends because they attend UNC, etc.). If I wanted to really analyze my social network, I would be able to distinguish the two. Don Park gets it somewhat close with the friendship circles, but that's still explicitly defining your relationships (albeit if you built a desktop app you could do drag and drop with thumbnails).
What I want to see is some sort of superproject (by Ph.Ds) that implements FOAF and uses some sort of program like Dashboard that constantly trawls your workspace and stores data.
For example, you could easily use the FOAF files to associate IMs/e-mails with existing users; if the program could also remember when these interactions occur, it could also remember the frequency of the contacts.
For example, if user "ABAlex " messages me in AOL Instant Messnger, a Dashboard-like program would pick up on this, grab my FOAF file and do a quick search of my friends to see if it can pick up on who ABAlex is. It would quickly find out that it's Borst ... and then store some data saying "Hey, Roy chatted with Borst on this date..." Everytime someone new messaged me or e-mailed me, the program would pick up on this and then implicitly remember that I am a "friend." Perhaps the engine might start using a simple frequency analyzer (# of interactions over a time span) to say that this person would be a "closer associate."
Of course, the downfalls of this? It's only limited to computer interaction; a lot of my 'real' friends don't bother with computers ... and how do you factor in phone conversations? Etc. etc.
Ok I'm not really making sense here. I have an exam in 5 hours and I'm up rambling about stuff like this again. Gotta go study ... gonna post this entry to the future. Cya in the future!
Comment with Facebook
Want to comment with Tabulas?. Please login.
Tallullah
tabulas
spaceinthewho
Thats where the processing power gets chewed up... cuz once you have a connected edge to someone, calculating your relationship to them is fast.
But its eliminating what I think is the coolest part of friendster... seeing how else you are connected to people!
Anonymous (guest)
Essentially a container widget which imbeds other widgets within itself. Thus allowing easy access to frequently used progs and freeing up desktop clutter. Some of my other GNOMEie friends and I are trying to lay down some base code...
will (guest)
MacDaddyTatsu (guest)
sal (guest)
sal (guest)
but apparently its what thousands (millions?) of people are into.