Skip to main content
Jump to: navigation, search

H5-11-03 B

Revision as of 11:05, 4 September 2006 by Paul.socialphysics.org (Talk | contribs)

(diff) ← Older revision | Latest revision (diff) | Newer revision → (diff)

We need an algorithm that will (given a Context c, and a time parameter t, and a (JUNG) graph g) update g adding and removing JUNG nodes and edges incrementally such that when complete g is an up-to-date graph that represents the contents of c at time t.

The algorithm (rough notes)

  1. Walk through the Context visiting every IDigitalIdentity (currently called IMembers) and every IEdge (probably should write a general purpose context walker) and simultaneously walk through graph g
  2. At every IDigitalIdentity and at every IEdge:
    1. Check to see if t lies within at least one of the DI/Edge's "LiveFromTo" properties
    2. If t lies within LiveFromTo then if a corresponding Vertext or Edge doesn't already existing in g then add it
    3. If t doesn't lay within a LiveFromTo property and is in g then delete the corresponding node within g

Added 12/06/05:

The above algorithm should be included in a package within the org.eclipse.higgins.ui plugin project.

An external JUNG-based visualization app (e.g. GUESS) could link in this .higgins.ui plug-in (and the rest of Higgins) and render the graph g (see [H5-11-04]). We will NOT attempt to integrate the rendering portions of JUNG or GUESS within the Higgins Demo app due to the fact that JUNG rendering is Swing-based (as well as for some open source licensing issues).

Back to the top