W3C home > Mailing lists > Public > www-style@w3.org > January 2010

Re: Making pt a non-physical unit

From: Henri Sivonen <hsivonen@iki.fi>
Date: Thu, 7 Jan 2010 15:10:22 +0200
Cc: www-style <www-style@w3.org>
Message-Id: <22937765-ADD1-44E1-BF42-30309E9C6038@iki.fi>
To: Jonathan Kew <jonathan@jfkew.plus.com>

On Jan 7, 2010, at 14:35, Jonathan Kew wrote:

> On 7 Jan 2010, at 11:39, Henri Sivonen wrote:
> 
>> On Jan 7, 2010, at 01:47, Jonathan Kew wrote:
>> 
>>> For those specialized use cases such as "life size" drawings, touch interface elements, etc., where a certain physical size really is intended - rather than a physical size that made sense on a typical computer monitor or a piece of paper, but should be appropriately scaled when viewing from 6 inches on a phone - I propose that we introduce a new keyword "true".
>> 
>> How would "true" units work with projectors and with big screens that are further away from the user than a typical display?
> 
> Just like the existing "physical" units ought to work according to the spec, AIUI, although the majority of implementations have departed from this.
> 
> It is of course fundamentally impossible for "true" or "physical" units to be accurately realized in all contexts; e.g., the case where a laptop is "mirrored" to a projector, so the same pixels are displayed on both, but at dramatically different sizes. Here, one of the displays - I assume the laptop - must be considered the "canonical" one for which physical units are correctly calibrated; the other is simply a zoomed projection of this.
> 
> (I'd expect projectors in general to be treated in this way, even if there doesn't happen to be a "normal" screen mirroring the projected display.)

I expect projectors to be treated so that 1in equals 96px. However, from the user point of view, there shouldn't be a rendering difference between a projected image of given true physical dimensions at a given true physical distance and a flat panel display of the same size at the same distance.

>> As far as I can tell, when Web content is displayed so that the display surface is further away from the user than a typical desktop screen, it's useful to make "physical" units relative to CSS px such that 1in equals 96px, since at distances further than user's reach, physical measurements of the display surface are irrelevant and only viewing angles matter.
> 
> Sure. So the "physical" units need to stretch so as to preserve the viewing angle - just as, for content displayed much closer to the user (think of a display built into a helmet), they need to shrink. This is why no "normal" content should be using true physical units. It may be convenient for some designers to express things in terms of points, cm, or inches, in the context of a typical computer display or a printed page, but I'm suggesting that ALL those units should be scaled together according to display size/viewing distance. What I dislike is the suggestion to treat "pt" differently from "in" or "mm" in this regard.
> 
> The proposed "true" units should never be used for any such content (and the nice thing about this, compared to defining *some* of the existing "physical" units as "truly physical" and others as "px-related", is that we can be confident NO existing content is using them). The "true" units are only useful for material - such as specific pieces of an application's interface - that is targeted at a known display size or similar context, not for Web content in general.

If browsers allow the use use of "true" units when media=screen (even if authors weren't supposed to use the units except for print and handheld touchscreen app UIs), we get back to the same problem as now with pt.

-- 
Henri Sivonen
hsivonen@iki.fi
http://hsivonen.iki.fi/
Received on Thursday, 7 January 2010 13:11:00 GMT

This archive was generated by hypermail 2.3.1 : Tuesday, 26 March 2013 17:20:23 GMT