Well I'm definitely not familiar enough yet, but agree that there are most probably un-welcome limitations.
Just as you (have) and I (am trying to) implement an interface... I don't see the reason yet, why one isn't provided for such a natural, symbiotic relationship.
Come to think of it, though maybe not practical as is, I believe a Model/View architecture should be at the Qt core itself. ...Everything in life thats not, in some way, a portrayal of chaos... is in a natural hierarchical topology/taxonomy. This holds especially true in computer science, me thinks.
Last edited by travlr; 7th February 2008 at 23:13.
Funny... Andreas Hanssen (vel Bitto) thinks the same way about QGraphicsView![]()
Graphics view and Model/View are conceptually similar, but implemented very differently. However, there is no reason you cannot write a QAbstractIemView that uses QGraphicsScene/View for visualization. It would be a considerable effort, but the results could be very nice.
Alternatively, since QAbstractItemModel is merely a wrapper around the underlying data, you can always have the QGraphicsScene access that data directly, bypassing model/view.
Any hints or tips on the current approach in your implementation?
I was thinking the same thing for a computer-aided design application: a MVC framework for the data, but at the same time, multiple representations (treeview, diagram using GraphicsView, regular 3D view using OpenGL etc...).
Choose one? Combine both?
Bookmarks