W3C home > Mailing lists > Public > www-style@w3.org > June 2007

RE: Proposal: font-size: auto

From: Paul Nelson (ATC) <paulnel@winse.microsoft.com>
Date: Mon, 11 Jun 2007 17:01:22 -0700
Message-ID: <49C257E2C13F584790B2E302E021B6F91395B184@winse-msg-01.segroup.winse.corp.microsoft.com>
To: Windy Road <tom@windyroad.org>, <www-style@w3.org>
The definition of “resolution independence” you are using is quite different than that which I would think about.


I would define resolution independence as the ability to render a page well without regard to the number of pixels per square inch available. Thus a 72 dpi, 96 dpi or 200 dpi rendering would show equally well.


It appears that you are defining resolution independence as the number of pixels that are available on the device. Thus, a hand held device might have text that is unreadable as seen by running your demo page in a browser window that is very narrow.


Hand held devices don’t need text that is smaller in size just because they have smaller screens. A user still needs 10-11 points of font size to view the text without having a lot of eye strain. In some scripts the font size needs to be even larger to be comfortably readable.




From: www-style-request@w3.org [mailto:www-style-request@w3.org] On Behalf Of Windy Road
Sent: Tuesday, June 12, 2007 3:59 AM
To: www-style@w3.org
Subject: Re: Proposal: font-size: auto


On 11/06/07, Paul Nelson (ATC) <paulnel@winse.microsoft.com> wrote:

	Scaling the font to fit this way would be based on a lot of assumptions. Do you want to do this by the average character width, the widest character width, require a monospace font be used? In Arabic text you have contextual character shapes which means the same character, like the beh, would have an initial that is very narrow and a final form that is very wide. What about scripts like Devanagari or Tibetan that have clusters or stacks and don't necessarily lay out linearly. 
	This type of computation would be *very* expensive and be more problematic for authors to know what to expect. For example, would one line have one size font and another line with other sets of characters be scaled to a different size to fit? 

The proposal would be for the formula to be very simple.  Generally in the form

  font-size = (width - padding-left - padding-right) / line-length 

No more and no less complicated than that. 

Paul, I'm not sure if you have been reading this thread so far, but the idea is to have in CSS a way of achieving the effect demonstrated at:


The demonstration uses the formula above.


Tom Howard

Received on Tuesday, 12 June 2007 00:01:28 UTC

This archive was generated by hypermail 2.3.1 : Monday, 2 May 2016 14:27:29 UTC