For iPad Retina (or iPhone Retina as well), why do we need to have two types of image sizes and add a suffix like @2x for the retina version?
Can't we just have one type of retina resolution images and for devices that do not have retina display, let the device handle resizing the image to smaller (non-retina) size?
Resizing takes time and memory. And resizing while retaining sharpness is hard to do. Which algorithm do you use? Bicubic or bilinear? "But this icon looks better when resized with other-algorithm-here!" How is iOS meant to know that the 1px border you drew is still meant to be a 1px border at half the resolution?