W3C home > Mailing lists > Public > www-svg@w3.org > September 2005

Re: Accessibility in SVG

From: Jon Ferraiolo <jonf@adobe.com>
Date: Sat, 17 Sep 2005 16:57:02 -0700
Message-Id: <>
To: Anne van Kesteren <fora@annevankesteren.nl>, Jon Ferraiolo <jonf@adobe.com>
Cc: www-svg@w3.org

At 10:29 AM 9/17/2005, Anne van Kesteren wrote:
>Quoting Jon Ferraiolo <jonf@adobe.com>:
>Thanks for clearing up the thing regarding coordination between the working
>The problem with unitless length "within SVG context" is that CSS does not
>solely apply to SVG, yet you want to use a single parser. It gets even more
>complicated in compound documents. Also, the CSS WG has its own ideas with
><number>. See for example
><http://www.w3.org/TR/2002/WD-css3-box-20021024/#the-width>. Therefore I think
>they consider it to be unacceptable for other WGs to extend their language in
>such a drastic way.
>The SVG WG could have required units in CSS context and let authors omit 
>them in
>the presentational XML attribute equivalents of those properties. This might
>still be a possibility as I heard multiple claims that no-one uses CSS 
>with SVG
>anyway and quite a few people using SVG rather saw CSS to be deprecated...

The above paragraph definitely is one option to consider as the CDF working 
group works through the integration issues between XHTML, CSS and SVG. 
However, it isn't my favorite option. As best I can determine, it does not 
help implementers at all (SVG implementers or XHTML+CSS implementers). The 
XML attributes in SVG were designed to accept the exact same syntax as the 
corresponding properties within CSS stylesheets or the 'style' attribute so 
that SVG viewers would only have to define a single parser for all possible 
ways of specifying a given property. In terms of the CSS cascade model, the 
XML attributes are defined as an author stylesheet where all of the 
specified values are assigned a specificity of zero. In my assessment, the 
sole merit of this option is that the CSS guys feel that they have achieved 
victory within a silly war of egos, but I do not see any benefit to 
implementers or end users, but perhaps I am not seeing a good technical 
reason why this option is a good one.


>Anne van Kesteren
Received on Sunday, 18 September 2005 00:06:32 UTC

This archive was generated by hypermail 2.3.1 : Wednesday, 8 March 2017 09:47:04 UTC