Re: Coments - last call draft

On Sat, 02 Apr 2005 02:27:52 +1000, Sean Hayes <shayes@microsoft.com>  
wrote:

> On CSS like styling, this has been talked about a lot and it was a
> deliberate decision to leave this out of DFXP, since it requires much
> more knowledge of the whole tree. AFXP on the other hand will have an
> applicative styling model which would allow user defined styles. DFXP
> would be compiled out of AFXP post user choice.

OK. So I am looking forward to AFXP. Like Al, I am concerned if DFXP goes  
to recomendation before AFXP has been at least through CR though, because  
I seriously expect some lessons to be learned there that are generally  
applicable.

> Having said that, user styling for accessibility is not always very
> successful, since to do it properly really requires knowledge of both
> how the content is designed as well as the specific requirements of the
> recipient. In an ideal world (which of course cannot exist since every
> recipient is unique, but might be approximated) the author/designer
> would provide a suite of style-sheets which can be swapped at a macro
> level. Such a choice mechanism might be provided through something like
> media queries.

User styling is generally reasonably useful for text, but not much good  
for other stuff. Maing it work better is partially a matter of authors  
providing alternative styles, but much more a matter of being able to  
describe the component parts of a document, and being able to apply style  
rules based on this. Which requires user agents to allow per-document  
style sheets, much as some browsers currently allow quite complex per  
document javascript configuration. And is a slow process, but at least one  
that can be done after the release of a specification.

The key point about CSS is that nothing stops me from shipping it to a  
user agent directly. Nothing stops me applying a CSS style sheet through  
the W3C Recommendation developed for that purpose many years ago and  
widely implemented. And nothing stops my user agent from allowing a user  
style sheet to be applied according to CSS. At the very least the spec  
should note this possibility.

> For a signing 'timed text', while work has been done both in animated
> avatars and codifying signing in a written form, to my knowledge this
> work has not been very successful to date from a legibility point of
> view. So I suspect that a video based approach is the only one that
> makes sense for some time. In which case SMIL might be the better
> starting point.

I think whether you use avatars or video SMIL is the better starting  
point. Which is why my coment concluded that an editoral note to this  
effect would be appropriate as a way of recognising the issue.

Cheers

Chaals


-- 
Charles McCathieNevile                      Fundacion Sidar
charles@sidar.org   +61 409 134 136    http://www.sidar.org

Received on Sunday, 3 April 2005 01:44:36 UTC