Re: Accessibility in SVG

Anne,
Just to set to historical record straight, since I was there at the 
beginning of the SVG WG:

>Also, I do not believe the SVG WG contacted the CSS WG in the beginning to
>discuss their extensions to CSS. If that would have been the case, some
>problems would probably never be part of the specification.

There was indeed very close contact between the SVG WG and the CSS WG, 
starting with the fact that in the first year or two of the SVG WG, we 
shared the same chair, Chris Lilley. The fact that SVG includes CSS at all 
is largely due to individual companies who had representatives on both the 
SVG and CSS working groups during the early stages of SVG, most 
particularly Adobe, Netscape and Microsoft, all three of which were strong 
proponents of defining SVG such that it could be implemented within a browser.

It just boils down to individual differences of opinion on technical 
matters, not historical lack of coordination. The long-standing question 
about whether unitless values should be allowed on CSS properties within 
CSS has been discussed countless times. The SVG WG believe that if the word 
"scalable" is part of your name, which takes you down the path of 
coordinate transformations such as scaling, which then pushes you into a 
requirement that you must define lengths and coordinates values which are 
relative to a current coordinate system, not an absolute measure such as a 
pixel or a millimeter. The CSS WG told the SVG WG had to put "uu" (for user 
units) at the end of any length value within a property specification. This 
would mean that every SVG length value had a ridiculous "uu" at the end of 
it, which serves absolutely no purpose except to satisfy mistaken purity 
within the CSS WG. Since every browser implementation supports unitless 
values on the HTML side for backwards compatibility with legacy content and 
has equivalenced them to "px" values, and therefore commercial CSS parsers 
already allow unitless values, and even if they don't, it is a trivial 
change, the SVG WG decided to not require "uu" at the end of every length 
value **within the context of SVG content**. The "uu" suggestion is absurd 
since it represents in incompatible change to CSS parsers which is thus no 
better than simply allowing unitless values, especially since almost all 
implementations support unitless values already. Certain members from the 
CSS WG continue to have sour grapes on this issue, but it was resolved 
years ago and now there are probably 50 different *commercial* 
implementations of the SVG language (let alone research implementations) 
all of which support unit length values. There is no way this can change at 
this point.

Jon

At 11:24 PM 9/16/2005, Anne van Kesteren wrote:

>Quoting Doug Schepers <doug@schepers.cc>:
>>However, I don't feel that the CSS WG has made any efforts at all to make
>>CSS generically work with other markup languages than HTML, so the fit with
>>SVG is rather poor.
>
>Why do you think so? It works rather well for generic XML languages. The 
>problem
>comes with languages where content and presentation are mixed, like with 
>MathML
>and SVG.
>
>Also, I do not believe the SVG WG contacted the CSS WG in the beginning to
>discuss their extensions to CSS. If that would have been the case, some
>problems would probably never be part of the specification.
>
>
>--
>Anne van Kesteren
><http://annevankesteren.nl/>
>

Received on Saturday, 17 September 2005 12:44:35 UTC