Finally, to make groups really useful, their names should have a smart behavior (like components) : defined pixel height at close-up levels, and % of screen size when zoomed out.
Moreover, if the user wants, it should be made easy to add Recievers in the left end of the group and senders at the right-end. Enforcing this has made my definitions worlds more understandable / de-bug-able / upgrade-able :
My vote for telepathy similar component.
I use it almost in any definition but there is a big problem on copy/past. If you do it with a large amount of telepathy elements every thing becomes a huge mess. A option for auto renaming a component when holding a key in a past command.
I never use the data viewer because I think it misses the point.
To debug / make progress in a definition with high order trees, it is often useful to see the contents of multiple components at once.
The panels are just fine, but their creation directly by double-click on component outputs would streamline the process.
I like the fact that the Data viewer adds the list of branches, but it also misses a great opportunity by not leveraging this list as a way to browse the data tree.
One could imagine clicking on a branch in the “Topology” list, and have the “Data” list scroll to that branch automatically.
Panels could also be much improved by the implementation of more precise scrolling, specially for large trees.
Once the elevator button is pressed and dragged, the scroll speed could depend on the distance between the panel and the mouse pointer, just like some video players have.
Pointer close to the panel = fast scroll
Pointer far from the panel (horizontally) = slow scroll