Originally Posted by Tallest Skil
Because it's going to be moot once we get Retina-style displays on the real computers?
Higher screen resolutions don't really help - if you increase the size of the screen enough, the bitmaps sampled for the UI will become pixellated. Similarly, if you scale the bitmaps to a size that can't be represented by whole pixels, you degrade the image quality. That's why Apple always doubles the x/y resolution of the iOS devices so that the upscaling is pixel-accurate.
The only solution is to migrate from bitmaps to vectors but it makes some things harder to do. You don't for example store rendered gradients, they are all rasterized at run-time so if you have a complex design, a vector format won't be able to reproduce it as easily as a bitmap. When it comes to a UI though, there's only so much complexity you need.
Apple have some support already. There are a few icons that are PDF vectors in the OS e.g:
If you QuickLook that file and scale it to fullscreen you can see it is sharp at any size. If you do the same with the Address Book icon, it pixellates but you can see the difference in detail as the AB icon has the leather pattern. It is possible to do very complex vectors using a gradient mesh but for a patterned surface, you'd still need a high density of points. Maybe Apple need to have a word with the guys at Pixar to come up with a tool to construct a best-fit procedural texture generator from a given input bitmap that executes on the GPU.
The easier route is to have hybrid graphics where the shape is defined by a vector and the fills defined by either vector or bitmap where the bitmap can be repeatable. The vector shape then acts as a mask for the bitmaps.