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

Re: Making pt a non-physical unit

From: Boris Zbarsky <bzbarsky@MIT.EDU>
Date: Thu, 07 Jan 2010 12:23:50 -0500
Message-ID: <4B4618A6.2040704@mit.edu>
To: Giuseppe Bilotta <giuseppe.bilotta@gmail.com>
CC: www-style mailing list <www-style@w3.org>
On 1/7/10 11:35 AM, Giuseppe Bilotta wrote:
> Fixing the px/pt ratio helps preventing that the same size as
> specified in pt come out as different design sizes depending on where
> the thing is rendered

I would hope not.  A UA should be choosing the design size most closely 
corresponding to actual rendered size, as much as possible.  For 
example, if you have a page that says it wants 12px fonts and it's 
zoomed to 2x in Gecko, Gecko will use the 24px font instead of scaling 
the 12px font.

> so you would
> actually be selecting the same design size for both screen and print.

That's a good way to get crappy output in one or the other (or both, of 
course).

>> Then there would be no unit at all to express "visible size" in CSS. That's
>> what px do right now, and imo is the one thing that's most important for
>> units used in CSS....
>
> I'm not sure what you mean by "visible size".

I mean precisely what CSS currently defines px as and what people 
actually perceive as "size" (modulo the automatic correction the eye 
does based on estimated distance): a given subtended angle of arc in the 
visual field.

> px as a single pixel makes sense for screen media, but it makes much less sense for print
> media, unless you plan to make 1px = 1 printer dot.

Sure.

> The proposal I'm putting my vote for is to make px=3/4pt=1/96in and
> then consider px the 'fundamental unit' in contexts where it makes
> sense (monitor, maybe projectors too) while taking pt (or in or cm or
> any other physical unit) as fundamental in contexts such as print or
> tablet devices where the actual physical dimensions have a higher
> priority.

Ah, I see.  But neither works for an iPhone, or an eyeglass display or 
contact lens display, in the way you defined them.  For a contact lens 
display at high enough DPI, equating CSS px to device px would give you 
pixels that are too small, and equating CSS px to 1/96in would give you 
pixels that are too big (all compared to the visual size of a pixel on a 
96dpi screen of 1998 viewed at typical arms-length).

-Boris
Received on Thursday, 7 January 2010 17:24:25 GMT

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