Re: Designs that zoom (was : Why reduce font size)

Thank you, I had the terms mixed up. The page also links to an
interesting discussion [1] in Mozilla's Bugzilla database. In a first
read-through, I find myself siding with "dcone (gone)" who reported
the bug, and who insists in the accompanying discussion that the
problem is somewhat but not completely solvable.

In the end, we're stuck with the limitation that user agents must
either make discrete estimates of pieces of a page, or try to
magically guess things from its overall structure. In the case of the
example over on positioniseverything.net , the user agent could
theoretically go, "hey, these are all part of a continuous block of
black, let's round some parts up and some down in order to make it all
fit". I believe this could be a working solution, but I also believe
that this is not an important requirement in a user agent.

[1] https://bugzilla.mozilla.org/show_bug.cgi?id=63336


On Thu, 03 Mar 2005 10:26:41 -0500, Ryan Cannon <ryan@ryancannon.com> wrote:
>  Pixel rounding is actually a phenomenon where browsers, when parsing
> relative measurements, are forced to round that width to a certain pixel.
> This usually depends on the precision of the browsers measurements ( i.e..
> rounding 10.555px to 11px, 11.6px or 11.56px, or even mistakenly 10px. Each
> of these can make a distinct visual difference, and it varies by browser.
> Positioniseverything.net has an excellent example[1].
>  
>      [1]http://www.positioniseverything.net/round-error.html
>  
>  This truly is a user agent problem and not a CSS problem, and may disappear
> as UAs become more precise. Perhaps CSS 2.1 or 3 should include a
> recommendation on how to handle pixel rounding?
>  
>  Carl Mäsak wrote: 
>  On Wed, 02 Mar 2005 11:55:59 -0500, Ryan Cannon <ryan@ryancannon.com>
> wrote: 
>  Some major problems were pixel rounding (which still isn't perfectly
> fixed), You might be referring to pixel perfection, in which case it will
> never be "perfectly fixed", as this is not the goal of CSS. Even the fact
> that different user agent manufacturers interpret the CSS specs differently,
> while not a "good" thing, will probably never go away completely. The below
> link points to a fairly sane discussion about pixel perfection and designer
> choices. http://www.alistapart.com/discuss/flexiblelayouts/7/ 
>  
>  
> -- 
>  
>  Ryan Cannon
>  Instructional Technology
>  Web Design
>  RyanCannon.com
>  (989) 463-7060
>  


-- 
Carl Mäsak
☎ office: +46184714159, home: +4618501392
✉ work: masak@ibg.uu.se, other: cmasak@gmail.com
☷ http://carl.masak.org

Received on Thursday, 3 March 2005 15:42:37 UTC