So we discussed in IRC yesterday that we might try to get the coordinates within the engine into a more easy-to-work-with state.
So here's a number of coordinate system, off the top of my head:
* Landscape coordinates
* Screen coordinates
* Viewport coordinates
* Object coordinates (transformed after Offset, DrawTransform etc.)
* Light Map coordinates
Most of those are translations, some of them scaling (aka zoom), and draw-transform could be an arbitrary transformation.
I would find it neat if we'd have a type-safe way to work on those - say have some new
How to get there, roughly: Have new unzoomed drawing routines in StdDDraw would be the start, accepting points of this form. Then we would define ways to convert
Does that sound workable?
So here's a number of coordinate system, off the top of my head:
* Landscape coordinates
* Screen coordinates
* Viewport coordinates
* Object coordinates (transformed after Offset, DrawTransform etc.)
* Light Map coordinates
Most of those are translations, some of them scaling (aka zoom), and draw-transform could be an arbitrary transformation.
I would find it neat if we'd have a type-safe way to work on those - say have some new
StdDDraw
functions accept StdPoint<Screen>
parameters and such - with the template parameter purely being used as a marker. We could then have conversion routines at the obvious places.How to get there, roughly: Have new unzoomed drawing routines in StdDDraw would be the start, accepting points of this form. Then we would define ways to convert
C4TargetFacet
to this form, with the plan to possibly transition to just passing around 4 points in landscape or screen coordinates respectively in future.Does that sound workable?
Powered by mwForum 2.29.7 © 1999-2015 Markus Wichitill