Rapid fractal compression remains hard. Some of Barnsley's techniques should be off patent now, or soon.
More important, if you abandon the vector form in favor of a bitmap, there's no going back -- information is irretrievably lost.
That's not a particularly useful question, because the direction of graphics programming is steered more by improvements in hardware than software. For example, I can remember writing and publishing a game for a programmable calculator, at a time when a "display" was a sheet of graph paper:
http://www.arachnoid.com/programmable_calculators/index.html...
My point is that I've personally witnessed huge changes in what people regard as cutting edge computer graphics, and most of the change is driven by hardware improvements -- GPUs, faster and cheaper computer power, display technology.
So to answer your question, I would advise that you pay attention to the hardware responsible for processing and displaying the images.