Are we optimising the wrong end.
I was reading about the google data compression proxy this morning when a specific line got me thinking:
“[the proxy] also optimises the perceptual quality of each image based on device screen resolution and pixel density of your device” (Data Compression Proxy)
So, why are we doing this on websites at the server end? After all a proxy has more information about the device, is easier to configure per user and moves the complexity to a single place
Moving this complexity downstream also removes the need for every site to implement their own approach to responsive images etc.
On a practical side, this is pretty interesting to me. I have always felt a little uncomfortable about providing retina images when i cannot reliably detect bandwidth and context.
I’m not sure how much i like the idea. But i do think its worth discussing.