2013/12/1 Nathan Hurst <njh@...1927...>:
I'm assuming that developing an estimate on rendering rewrite will take some time. Is this correct?
It's possible that Krzysztof Kosinski or his summer of code student would have a fairly good idea of what is involved for this.
The current rendering pipeline is in fact fairly well suited to animation, because it goes out of its way to only compute / draw the things which are necessary. If we are sticking to Cairo, it only needs some optimization and minor additions.
A fairly big rewrite would be required if we wanted to support pluggable renderers. For instance, we could try adding alternative renderers based on Skia, Mozilla Azure, OpenVG or OpenGL (Cairo's OpenGL backend will not give us maximum performance, because the Cairo API does not correspond very well to how OpenGL works) or implementing some export formats in this way.
Note that I do not know enough OpenGL right now to implement an SVG renderer in it, or much OpenGL at all, though I could perhaps propose this as my Master's thesis in CS.
Regards, Krzysztof