?

Log in

No account? Create an account

Previous Entry | Next Entry

Progress

Ended up writing about 2,000 lines of code today even though my Perl is a bit rusty, but now have working code. It can crawl URLs looking for FOAF, XFN, and hCard relationships/markup and dump the nodes and edges (storing types) into a database. I like pictures so here are a few graphs (of fully public data):

Myself:
http://pics.livejournal.com/daveman692/pic/001se7hh

Joseph Smarr:
http://pics.livejournal.com/daveman692/pic/001sc6zr

Mark Pasc:
http://pics.livejournal.com/daveman692/pic/001sdeqe

Comments

( 3 comments — Leave a comment )
andyksong
Aug. 31st, 2007 08:30 am (UTC)
Great job, David
You make it really fast!
shechtman
Aug. 31st, 2007 03:50 pm (UTC)
These look very promising. But what about Facebook?..
fallenpegasus
Aug. 31st, 2007 06:16 pm (UTC)
I'm sure that the social networking group has already talked this out, but I think that 80% of the initial problem can be solved by designing an "auto discovery" protocol, similar to the ones for OpenID and for Atom, to discover a social network user's vCard and FOAF data. And to put into the FOAF and vCard the references that the user gives the site about other sites they are on.

Then a script like the one you just wrote, running in background on each site, can easily constantly discover relationships, and then occationally tell the user "ive just discovered the following relationship between you and user foo here. do you want to friend that user here too?"
( 3 comments — Leave a comment )