I just watched the GraphStream
demo<>and it looks
fantastic. What does it take to integrate it with MASON?


On Fri, Apr 19, 2013 at 11:04 AM, John McManus <[log in to unmask]>wrote:

> As an option, I have found that the newer GraphStream plays very nicely
> with
> Mason. It does not have as extensive a list of metrics as Jung yet, but the
> graphics are superb!
> Note that it did not play well with BIRT.
> Cheers!
> John
> -----Original Message-----
> From: MASON Multiagent Simulation Toolkit
> [mailto:[log in to unmask]] On Behalf Of Sean Luke
> Sent: Wednesday, April 17, 2013 11:52 AM
> To: [log in to unmask]
> Subject: Re: MASON with JUNG
> So a bit of history.   Back in the early days of MASON development, we put
> together a social networks package due to an internal request.   It's
> called
> 'socialnets' and is part of the contrib packages right now.  That package
> isn't used much and hasn't been tested much.  But it's pretty well done i
> think, and with some polish maybe we should roll it into the rest of MASON.
> 'socialnets' has a lot of analysis facilities but doesn't have any
> visualization: it builds on MASON's network package and so its social
> networks may be visualized using MASON's network facility.  Which is
> moderately primitive.
> Maciej Latek, here at GMU, wanted to use JUNG instead, and decided to write
> some code which bridged the two a while back.  That's the Jung package you
> see on the MASON website, and it's the 'jung' package in the contrib
> directory.  It's not been updated in quite some time, and so no doubt is
> suffering from bitrot as MASON gets updated.  Also Jung isn't known for its
> speed.
> Hooking MASON up with other people's libraries often entails a bit of
> hassle
> because these libraries often aren't particularly well designed, and are
> rarely separable, serializable, or self-contained.  It's annoying: they
> don't think about interoperability with other toolkits.  For example, we
> had
> to bug the JTS folks to get them to add serialization so GeoMASON would be
> serializable like a proper MASON toolkit.  JUNG is typical of this
> unfortunately.  It doesn't help that JUNG hasn't been updated for three
> years.
> So anyway with some tweaking you should be able to get the JUNG toolkit
> working again with MASON -- we haven't deviated that much.  Also Richard,
> if
> you'd like to add some contributions to the JUNG contrib directory, or
> indeed take it over, let me know.  And last, try the socialnets package.
> It'd be nice to have people banging on it and finding bugs or making
> improvements -- it's a nice little package I thought.
> Sean