Re: [CSS21] 4.3.2 Lengths (reference pixel?)

On 2010/12/13 11:10 (GMT-0800) Tab Atkins Jr. composed:

> more than
> one browser, at least, has done the "1in = 96px" thing for a long
> time.

AFAICT, that's the only way WebKit has ever been or could be. Presto defaults 
to it, but can be configured to use a selected DPI. For several years the 
Geckos have floored at 96, using actual when higher. Previously they simply 
used whatever the DTE was configured to or assumed to be, as IE7 and prior 
did. I don't recall whether KHTML floors at 96 or not, but otherwise it works 
like non-beta Geckos.

> So there never was a way to do it "correctly" because the physical
> units never were truly physical, in practice.

That's true only in the simplistic worlds of OS/2, Mac and Windows. The 
ability to match display device to desktop configuration, resulting in 
accurate[1] absolute dimensions at the display surface, has been a part of 
X11 on Linux at least as far back as my experience with it goes, either by 
default, or as an option. A number of recent Linux distro releases assume 96 
DPI, while others make no assumption, favoring accuracy, usually by using 
display dimensions and resolution provided by EDID to calculate DPI 
automatically and accurately[1]. In any event, not only is it possible for 
absolute units to be accurate on Linux, it's very common.

[1] +/- roughly 1 device px
"The wise are known for their understanding, and pleasant
words are persuasive." Proverbs 16:21 (New Living Translation)

  Team OS/2 ** Reg. Linux User #211409

Felix Miata  ***

Received on Tuesday, 14 December 2010 07:16:21 UTC