W3C home > Mailing lists > Public > www-style@w3.org > March 1998

Re: browser, rez, OS issues

From: Todd Fahrner <fahrner@pobox.com>
Date: Mon, 16 Mar 1998 16:22:06 -0800
Message-Id: <v03102803b1336a880a12@[206.245.203.103]>
To: George Olsen <golsen@2lm.com>, www-style@w3.org
George Olsen wrote (3:32 PM -0800 3/16/98):

" If you've designed *everything,* including images, with relative measures,
" then yes your page can scale and you don't need to suggest a viewing
" resolution. However, using relative sizes for images can cause undesired
" results as the images are resized by the browser -- gifs in particular
" don't scale well. (Which points out the crying need for a W3C standard for
" vector-based graphics.)

Agreed. If line/spot-color art were typically webbified in a more
approppriate format, eventually resolution wouldn't matter much. But
because scaled GIFs look bad, nobody seems willing or able to implement the
CSS1 Recommendation WRT pixels (see below). It's chicken-and-egg.

" Unfortunately, typographic sizes *are* still dependent on the OS (i.e.,
" type on Windows will be about a third larger than on Mac -- don't know what
" Unix does) unless you've specified the type size in pixels. However, even
" though pixels would be the best choice for resolving the Mac/Windows
" problem, using pixel dimensions causes even bigger problems because CSS-1
" provides no way of specifying an implied resolution for different media.

Almost. CSS1 defines a "reference pixel" as a degree of visual angle
equivalent to 1/90" at arm's length, with the admonition for UAs to scale
elements should the actual visual angle of a pixel vary "significantly."
That's a tall order (does your browser look back at you?), but now that one
vendor has gone ahead and made the browser part of the operating system,
they might eventually get hold of the information necessary to scale pixels
properly. The other vendor scales pixels based on a virtual pixel of 1/120"
when printing. [doh!]

" Consequently, when 30-pixel high type is printed out on a 600 dpi printer,
" it's a faction of a point high.

This was a bug affecting IE3 only.

" However, there is a relatively easy solution, at least for typography. Set
" up duplicate site-wide style sheets for each OS and adjust the type size
" accordingly.

I confess that I'm guilty of similar tricks, but I try to remind myself
that Windows doesn't have a fixed reckoning of a point. It has a default of
1/96", but that's overridable, and as display resolution edges up over the
years, it will routinely be overridden.

I haven't tried specing height and width of images in points (IE4 only),
then remapping the pixel value of a point and seeing if it scales
appropriately. Even if it worked, though, it would be a betrayal to my Mac
brethren to use it - we can't mess with how points render in most apps.

The CSS1 Recommendation says that absolute units like points are only
useful if you know the physical properties of the display. That's a mystery
I'd rather not have to investigate with the feeble tools currently at my
disposal, nor ever dismiss with some half-baked guess.

" Then you can simply do an browser or OS check and call up the
" appropriate style sheet. I suppose you could also do check for monitor size
" and set up CSS-P style sheets based on that, but that's probably equivalent
" to simply rebuilding the page.

What does monitor size have to do with window size? And what about all that
chrome? <g>



Todd Fahrner
mailto:fahrner@pobox.com
http://www.verso.com/agitprop/

The printed page transcends space and time. The printed page, the
infinitude of books, must be transcended. THE ELECTRO-LIBRARY.
	- El Lissitzky, 1923
Received on Monday, 16 March 1998 19:15:13 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Monday, 27 April 2009 13:53:54 GMT