Hi Eytan, thanks for the reply. The approach
that you outline sounds interesting, but
unfortunately, as you've already guessed, the
main requirement is for the subgraph nodes to
be displayed as a single unit within a bounded
box, with incoming edges pointing into the box
(i.e. into the whole subgraph, not to any particular
vertex of it), and with outgoing edges coming
out of the box as a whole. Subgraph boxes might
also need distinct labels. And this capability
of course would need to be recursively available,
to an arbitrary degree of nesting (within practical
visualization limits). If you do plan to
implement something like this soon, please let
me know, as I would very much prefer to use
Guess than any other framework I've seen so far.
Thanks!
-- K.
--- On Tue, 6/17/08, Eytan Adar <eyta...@gmail.com> wrote: