W3C home > Mailing lists > Public > www-style@w3.org > September 2006

Re: Sizing

From: Felix Miata <mrmazda@ij.net>
Date: Sun, 17 Sep 2006 09:27:30 -0400
Message-ID: <450D4D42.9050706@ij.net>
To: www-style@w3.org

On 06/09/17 12:28 (GMT+0100) David Dorward apparently typed:

> On Sun, Sep 17, 2006 at 01:07:32PM +0200, Axel Dahmen wrote:

>> What is rendered on screen always varies depending on the screen
>> resolution.  Units like "pt" are never displayed in their correct
>> sizes.

> Most systems are not calibrated correctly so the system can't convert
> points to pixels accurately (there may also be browser bugs
> involved). I don't see how any change to the CSS specification could
> work around this problem.

The calibration error was traditionally and nearly universally very
significant until several years ago. Before the turn of the century,
640x480 was the most common resolution, which on a typical 13" viewable
CRT display of the time was only  about 61 DPI. 800x600 on 14" viewable
was about 71 DPI, and 1024x768 on 16" viewable was about 80 DPI, all
using an OS-assumed system DPI of 96.

With the spread of flat panel displays and use of manufacturer's
recommended native resolution on them in most cases, DPI error has
become small enough to be arguably inconsequential. The ubiquitous OS
still assumes 96 DPI, but manufacturer's of the highest resolution
displays tend to either by themselves or by inducing the user move that
assumption to 120, thus reducing or eliminating significant error.

The upshot is the statement that 'units like "pt" are never displayed in
their correct sizes' while once true is no longer the case if
insignificant error is disregarded, as a study of
http://mrmazda.no-ip.com/auth/Font/fonts-pt2px-tabled.html should show.

>> Similar for printing: "px" unit rules are never printed in the
>> printer's resolution.

> They shouldn't be:

>    Pixel units are relative to the resolution of the viewing device,
>    i.e., most often a computer display. If the pixel density of the
>    output device is very different from that of a typical computer
>    display, the user agent should rescale pixel values.

>      -- http://www.w3.org/TR/CSS2/syndata.html#value-def-length

What do "very different" and "typical" mean? The spec apparently assumes
the later to mean 90 DPI, but what of the variation? 10%? 25%? 50%?
100%? And why 90 rather than 96 or 100? My 20" 4:3 flat panel display
has a native DPI of 87.5. Is that typical, or very different?

>> The new "scaling" rule should only allow
>>   * percentage values
>>   * a factor to set a fix scaling, using length units:
>>       something like
>>          "300pt" => 300 pixels equal one Point
>>          "500cm" => 500 pixels equal one Centimeter

Web users would be better off if the px unit were removed from the spec
entirely. The px unit seems to be useful primarily for two things:

1-making it easy for authors to design non-adaptable magazine pages for
hosting on the web
2-making text tiny, disregarding user text size preferences

Most people would be better off with a spec that induces most authors to
use relative units entirely, in most cases, relative to the user's font
preference and/or viewport size. Authors would decide how everything
should fit together proportionally, but users would determine how big it
should be overall. After all, most users are using _personal_ (and thus
personally customizable according to taste and requirements) computers.
http://mrmazda.no-ip.com/SS/bbcSS.html illustrates the sorry state of
what the web is like now as a result of the CSS px unit.
-- 
"Wisdom is supreme; therefore get wisdom. Though it cost all you
have, get understanding. Esteem her, and she will exalt you;
embrace her, and she will honor you."		Proverbs 4:7-8 NIV

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

Felix Miata  ***  http://mrmazda.no-ip.com/
Received on Sunday, 17 September 2006 13:27:13 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Monday, 27 April 2009 13:54:46 GMT