I really wanted to start this by pointing out that Ambient has been stuck at 1.45 for about 20 years, only to find out that it has been upped to 1.47 at some point..... well duh....

Ignoring the version number (which I do think should ge to 2.0 if not equal the MorphOS version it comes with) here are some thoughts about it's future. Some of this may never be done, other parts only as optional 3rd party effords and some may be decided to be bad ideas from the start.....

I am leaving Panel features out of this as there is an extra thread for it already.

- D&D/AppWindow
MUI D&D only works within one app while the existing AppWindow API won't give any feedback on objects being dragged over such a window.
Current (hacky) solution for Panels is setting the userdata to a fixed value and than adding code in Ambient that generate special messages your dragging something over such a window instead of just changing the mousepointer.
Future solution: AddAppWindow() does have a taglist argument that AFAIK isn't yet used anywhere, so defining WBAPPWINDOW_MouseReport and AMCLASS_MOUSEMOVE/AMCLASS_MOUSENTER/AMCLASS_MOUSLEAVE (+AMCLASS_DROP maybe) should do that trick in a proper way.

- Prefs and prefs files
Atm I do have a "panel2.library" in PanelApp that mostly deals with those for the panels, but it might be a good idea to move that into a "Ambient.library" or "prefs.library" giving access to all such files
PanelPrefs has been moved to an mcc like extenal class that will eventually be called from Ambinet (hence the need for handling prefs file via a library) all other pref-panes could also be made into such classes. Would lead to a slight reduction in binary size and memory consumption (when prefs are not used) with no real changes otherwise.

- Listers
These could also get some support for 3rd party extensions/subclasses. Tabbed? With a listtree on the side? Coverflow?

- Search
Make that a proper lister, so you can D&D files out it