W3C home > Mailing lists > Public > www-style@w3.org > April 2006

Re: Downloadable fonts and image replacement

From: Chris Lilley <chris@w3.org>
Date: Fri, 28 Apr 2006 16:50:35 +0200
Message-ID: <1013949836.20060428165035@w3.org>
To: "Rijk van Geijtenbeek" <rijk@opera.com>
Cc: www-style@w3.org

On Thursday, April 27, 2006, 4:02:30 PM, Rijk wrote:

RvG> On Mon, 24 Apr 2006 16:56:11 +0200, Bert Bos wrote:

>> The CSS Working Group has been discussing downloadable fonts ("web
>> fonts") and the interaction of font choice and image-text replacement
>> recently.

>> Some of the issues under consideration
>> --------------------------------------
RvG> [..]
>>       - Accessibility considerations for image-text replacement

RvG> I'd like to mention the bad habit on many Indic websites, where
RvG> pages in Bengali and Mayalayam (and some other) scripts are send to
RvG> the browser as Latin-1 or as charset=x-user-defined, and the
RvG> non-Unicode fonts needed to make this work are offered as a
RvG> separately downloadable ttf or as eot with the @font-face styles.
RvG> Demo: http://www.jugantor.com/demo/ .

RvG> The spec might not be the most useful place to fight this kind of
RvG> web authoring, but it should be made very clear that the usability
RvG> of websites should not depend on the availability of a certain font.


That is a long standing problem. It used to be much worse, eg it was common even for Greek and so on in the early days.

Because of this, the appreciation of the difference between character encodings, Unicode code points, font encodings, and bytes on the wire at W3C started with the old Font activity. W3C rapidly adopted the ISO character-glyph model, based on contacts at Unicode conferences.

Not just rendering is affected, but also such basic things as searching. Imagine having to type in several different types of latin garbage into Google to cover several popular font hacks (each with different 8-bit cmaps) just to find information - and of course the results page is unreadable in that case, combining different hacks in the one page.

This Unicode-based character glyph model has since spread to be general W3C policy and is mentioned in the Architecture of the World Wide Web and in the CharMod Fundamentals.

http://www.w3.org/TR/webarch/#archspecs
http://www.w3.org/TR/charmod/

For more complex scripts, there is an outstanding issue of deployment of the language support, such as complex character to glyph mappings for some Indic scripts. This capability was added late to some OS (such as XP) and is not available in earlier versions (such as Win98 or macOS 8..9) which are still deployed in considerable numbers in developing nations. For some languages, even latest versions of XP, MacOS X,latest Linux distros do not have suitable support.

For those languages, latin-1 hacks or even gif-as-web-page hacks are still deployed.

Over time the deployment and applicability of those hacks is shrinking.



-- 
 Chris Lilley                    mailto:chris@w3.org
 Interaction Domain Leader
 Chair, W3C SVG Working Group
 W3C Graphics Activity Lead
 Co-Chair, W3C Hypertext CG
Received on Friday, 28 April 2006 14:50:43 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Monday, 27 April 2009 13:54:44 GMT