Lazy initialization

[akevalion]

hello,

small demo of the kind of things we expect to be able to do:
https://www.dropbox.com/s/cu7lq78qlzd75im/Kazam_screencast_00000.mp4

This is a medium model of all packages/classes/methods for Athens, with
all packages/classes/methods that they use (Object, Collection, ...)

I created a visualization very straigth forwardly (code attached):

• get all packages from the model -> create nodes,
• get all these packages' classes -> create nodes
• get all these classes' methods -> create nodes
• then find all invocation links between the method and create
  #dependencyTo: the corresponding nodes

First, it's a bit long to get the visualization to open (15 / 20
seconds)

Second, when it opens, we don't see anything meaningfull, it would be
nice to be able to say that we want to open with the nodes collapsed,
and possibly without any link drawed

And again, when I collapsed the node, it is very long (another 15 sec)
For indication, I collapsed the node and immediatly dragged the window.
You can see how long it takes for the window to be actually moved

[NicolasAnquetil]

I deleted the video from my dropbox.
Link above is no longer valid
I restaured the video, new link is: https://www.dropbox.com/s/l4wf1sa42yrfuod/Kazam_screencast_00000.mp4

[NicolasAnquetil]

Hello reviving this issue
With @LABSARI We wanted to see 2 packages in a big model.
Starting the visualization is very long because it computes all the hierarchy before displaying anything.

To be more efficient, the visualization should only compute the "next level" : for each displayed node, compute its children (to know if there are some or not).

And only when the user expands a node, its direct children are shown and their direct children are computed (again, to know if they have some or not).