From: Anders Sandberg (asa@nada.kth.se)
Date: Mon Aug 05 2002 - 09:10:11 MDT
I have written a small application for displaying general graphs in 3D,
laying them out with a simple force model. Very fun exercise in XML and
OpenGL, and I plan to run it in our local Cave.
As one experiment, I looked at the list archives for July 1 and created
the graph of replies (person A is linked to B if A replies to B in at
least one post); a simple screenshot can be seen at
http://akira.nada.kth.se/~asa/extrojul1.gif (it is much clearer when it
is rotating, but you get the idea). Very cool, but making it by hand was
somewhat cumbersome. Now, being a datajunkie, I would love to make a
larger graph of (say) last week's postings, but I really need some kind
of filter that automatically trudges through the list archives
outputting lines of who each post is a reply to (if any), so that I can
turn the traffic into a graph (size of a node could be number of posts,
while color could be a function of mean length, number of original posts
and (say) the crackpot index :-). Is such a filter possible, or are
reply formats so erratic that one has to do it by hand?
Does anybody out there have other fun graphs to visualise? Citation
graphs, neural networks, powergrids? My app can read a subset of GraphML
(http://graphml.graphdrawing.org/), essentially a list of nodes (<node
id="foo" color="1 1 0"/>) and edges (<edge source="foo" target="bar"/>)
with extra data attributes.
-- ----------------------------------------------------------------------- Anders Sandberg Towards Ascension! asa@nada.kth.se http://www.nada.kth.se/~asa/ GCS/M/S/O d++ -p+ c++++ !l u+ e++ m++ s+/+ n--- h+/* f+ g+ w++ t+ r+ !y
This archive was generated by hypermail 2.1.5 : Sat Nov 02 2002 - 09:15:53 MST
