Re: graphs

• To: "Steven Smith" <sas@xxxxxxxx>, "Oscar Stiffelman" <oscar@xxxxxxxxxxxx>, "Bill Hibbard" <hibbard@xxxxxxxxxxxxxxxxx>
• Subject: Re: graphs
• From: "Manuel Simoni" <rainhard.kofler@xxxxxxxxx>
• Date: Fri, 25 Jan 2002 14:54:12 +0100

Steve wrote:

> Some of the graphs (especially if I take them over time) are dense enough
to
> suggest connectivity matrix representations and especially if I use a
"sparse
> matrix" package.  Does anyone here have anything to suggest in that realm?

Java graph packages (or packages that contain graph support) I have stumbled
over are:

http://OpsResearch.com/OR-Objects/

http://www.cs.brown.edu/cgc/jdsl/

http://www.infosun.fmi.uni-passau.de/GTL_java (I think uses native code)

C++ http://www.boost.org

> CAIDA (www.caida.org) has built an interesting layout/manipulation tool
known
> as Walrus (www.caida.org/tools/visualization/walrus) which implements a 3D
> hyperbolic metric for layout and investigation which maintains excellent
> scaling properties.  The early work was by Tamara Munzner but this is the
best
> example to date, that I know of  for handling the "detail in context"
problem
> around graphs.

Walrus looks very interesting, I have just requested a copy.

> It seems possible that we need a whole new base data structure that is
parallel
> to the ones already in Visad along with some conversion methods...

Yes.

Manuel

• References:
  • graphs
    • From: Oscar Stiffelman
  • Re: graphs
    • From: Manuel Simoni
  • Re: graphs
    • From: Steven Smith