On Wednesday 2012-04-18 08:32 -0700, Ryosuke Niwa wrote: > On Wed, Apr 18, 2012 at 1:26 AM, Kang-Hao (Kenny) Lu < > kennyluck@csail.mit.edu> wrote: > > > (12/04/18 15:12), Aryeh Gregor wrote: > > > As another idea: always output with the 'rem' unit when "styleWithCSS" > > is on. This is as you said, relying on the 'font-size' of the root > > element being fixed, but I think this is better than adding a value > > that's mostly useless. > > > > That won't work because rem isn't supported by old browsers and mail > clients. In editing, simply generating contents that can be rendered by > yourself isn't good enough. It needs to be renderable by other browsers. So as has already been pointed out, this seems to defeat the argument that an xxx-large value would be useful. One could in theory imagine a new unit, say 'uem', that's like 'rem' except it's based on the user's default font size preference (i.e., doesn't include author changes that apply to the root element). However, I tend to think that's probably not the best of ideas, since I think the whole idea of user font size preferences has failed, for two reasons: * First, on many pages, authors specify fixed font sizes and ignore the user's preferences. This means that when the user tries to adjust the preference by looking at such a page, the adjustment appears to do nothing. * Second, many other pages make assumptions about the user's default font size being a specific value or within a narrow range of values. On these pages, changing the user font size preference from its default value can cause overlap, hidden content, or other problems. I tend to think that the current font size preference mechanism should be replaced by: * Making the current default values of font sizes the only values. * Honoring any user font size preferences (whether they're fixed values, minimum constraints, or range constraints) by zooming the entire page (i.e., displaying the page at zoom factor f involves laying out the page as though the viewport had been scaled by 1/f and then scaling the resulting display back up). This approach would make the xx-small to xx-large keywords equivalent to pixel font sizes. -David -- 𝄞 L. David Baron http://dbaron.org/ 𝄂 𝄢 Mozilla http://www.mozilla.org/ 𝄂Received on Tuesday, 24 April 2012 18:21:02 UTC
This archive was generated by hypermail 2.4.0 : Friday, 25 March 2022 10:08:15 UTC