thanks
I already started a little bit on getting the basics for context rendering, I'm useing what I call omnipoints to represent 99.9% of everything we are going to need, the idea I have is that there is a 2d vector of these and the GL canvas doesn't need to know anything about them, the editors will simply pass a few flags about what should be rendered the GL canvas modifies the value for one or more of the points, then sends the vector back to the editor which applies the changes. the editors already have some (sloppily written while I was waiting for bug reports) get and set functions for omnipoints, mostly what needs to be done interface wise is get and set omnipoint functions for the GL canvas, and get and set functions for which point is selected (the function would have two ints).
the biggest pitfall in this system is that the number of omnipoints in any part of the vector absolutely can not be changed by the GL canvas, and the editor absolutely must make sure it updates the GL canvas with a new omnipoint vector anytime any data (especaly new points/banks) is altered. so we will need to basicly have a 'start edit' mode when the user clicks on the GL canvas with a key pressed (or something) were the data in the editor is transfered (copied actualy) to the GL canvas (or maybe not sence the editor should be keeping the GL canvas up to date) and then the GL canvas makes changes to the data and the editor can not make changes to it's own data untill after edit mode is ended and the data is transfered back to the editor. I'm prety sure most if not all edit modes would only be operating on one (set maybe, if we allow multable selection in the tree) of data at a time, so this probly shouldn't be as monolithic as it might sound, I'm just going into detail on something I've only partaly worked out.
but there is still a fair bit of cleaning to do with what we have before we get side tracked on that, if you want to work out a way of efficently drawing a set of spheres of a set position and radius with a specific color (preferably useing vertex and index buffers), that would probably be a good use of time in regards to this. there will be three seprate colors for this to deal with, unselected, selected row, and selected item, I'm sure the meanings of these are obvius. I think geodesic spheres in wire frame and lightly alpha blended, if posable, would be cool.
