kwando : that kind of change you'll be able to do directly in the editor code, no need for modules
Ganon : nothing wrong in itself, it's a good approach on big projects.
But for a feature-oriented modular approach to be useful, you need each feature module to be independent of the rest, or have a few explicitly defined dependencies : if you edit, enable or disable one module, you don't want it to have a potential impact on all the other ones nor on the common UI.
This implies specifying precisely the way a module can interact with other modules or with the core application, which is a loss of flexibilty compared to the current way in which I just plug code where I need it. (I would probably need to revamp everything to go for an event-driven model in the player if I wanted it to be modular but still reasonably flexible.)
This also requires a fluid UI : a UI which can display an unlimited number of information (you don't want to rethink the UI every time someone adds a new module) without knowing their nature. Hence a loss of flexibilty on the UI design.
As a (bad) example, think of it this way : currently, the editor knows exactly what set of features it has to expose for a frame, so I can place them exactly as I wish. Screen on the left, music and sound on top of each other, etc.
If you take a modular approach on these features, it means the editor main interface should be able to load and handle abstract modules without knowing what they contain. The editor just knows that it has to render something for each module it loads, but delegate the rendering to the module's own code.
So, if screen, sound and music are three modules, how would the editor know that the sound and music modules are both small enough (and related enough) to be fit together ? And if I edit the music module to add an option, the corresponding area in the editor will be bigger : how will the UI adapt to that, so that the sound module is not hidden ?
Of course, it's only a silly example - you're not suggesting to split apart frame components as features, I believe. But I hope you get the idea.
While I use a modular architecture in the V6 code, it's oriented towards technical components (some modules for screen rendering, some modules for action management, some modules for editor interface, etc) : going for a well designed feature-oriented approach would be considerably more complex, and probably not needed given the size of the project !
Sing : yes, all of this would be possible improvements