What I don't get is why the 2x image would ever have a different viewport size than the 1x. It seems like something no one would ever use, as the layout would change depending on the user agent's image resolution preference.<p>In any case, I think both of these solutions are woefully incomplete, and quite inefficient. Why do we have to maintain lists of file paths for each resolution of every image when we could take advantage of file naming conventions in a systematic way and avoid all that work and complexity? Sets of images invariably share almost exactly the same file name and the variations are largely standardized for any given web page (all 2x images have "_2x", "@2x", "@2", etc.); surely we can use that to our advantage.