From: Dr. P. H. <pl...@in...> - 2004-11-15 18:54:13
|
Francois, Martin, some weeks ago you were discussing the issue of=20 black-box shapes, grouping of shapes and perhaps collapsing and expanding of grouped shapes. I think that would be a really useful feature. It=20 could help a lot to overview complex graphs by=20 a kind of hierarchy. Here are my 2 cents: Perhaps it could be implemented by a "group-shape" with a boolean variable "expanded". All grouped shapes would need a reference to this parent group-shape. The reference can be null, if the=20 shape is not grouped. So, when it comes to paint, the shape would branch: If no parent exists or the parent is expanded, it=20 is painted as usual. If not, it is not painted at all. On the other hand, the group-shape is only painted, when it is not expanded.=20 The group shape would need a special kind of connector=20 for each connection between an included shape and a not=20 included shape. These connectors on the group-shapes=20 need to behave like the subpoints on the connections,=20 so they won=B4t disturb the logic. Of course, there will be the issue with the positions=20 of the shapes, that Martin mentioned in his post. Perhaps the automatic layout could help there. What do you think? Greetings=20 Helge |