1) 
Good point here "Apple can afford it as they control both HW and SW".

Let us take an example: say we have 14" and a 15" laptop with same physical resolution. If the logic is simply to scale a factor 2 if DPI exceeds a certain fixed threshold, it may happen that this threshold is exceeded on the 14" screen and not on the 15" and icons, controls, text etc.will be printed to the screen almost twice as big on the 14" screen as compared to the 15".  This makes little  sense to me. The small one may have a better quality monitor and could in some cases actually better be able to cope with unscaled display than the bigger one. An one person may be able to (and may prefer to) ane another person may not.


2)
"The web has a concept of device pixels vs css pixels as well".  Does anybody know how Chromebooks handle it? That could bring a new perspective into discussion.




-- Peter

On Mon, Jan 12, 2015 at 12:55 PM, Andreas Nilsson <lists@andreasn.se> wrote:
On 2015-01-12 12:23, Michael Stahl wrote:
http://robert.ocallahan.org/2014/11/relax-scaling-user-interfaces-by-non.html is there something particular to HTML/CSS that makes this work better for the web than desktop apps, or do web users have lower expectations of rendering quality, or what?

The web has a concept of device pixels vs css pixels as well. It allows for websites to look nice by default on hidpi screens.
http://www.smashingmagazine.com/2012/08/20/towards-retina-web/
Zooming of websites is a nice bonus to have, but doesn't solve the same problem.
- Andreas



--
Hilsen / Regards

Peter Laursen