W3C home > Mailing lists > Public > w3c-wai-gl@w3.org > January to March 2001

Re: class definitions and grouping - example 1

From: Marja-Riitta Koivunen <marja@w3.org>
Date: Mon, 12 Mar 2001 13:51:42 -0500
Message-Id: <>
To: Kynn Bartlett <kynn-edapta@idyllmtn.com>, w3c-wai-gl@w3.org
I was trying to see a bit further than what we may or may not have today as 
there is always that chicken and egg problem.

Many things maybe cumbersome now, but when user agent let users to have 
several ways to browse the pages and the elements on the pages, it could be 
different. At least at one point the UA guidelines had all kinds of 
browsing mechanism that should be offered to the user. Maybe a site could 
even provide some accessibility friendly metadata of the class or other 
conventions they are using.

I don't know how you want to use XML elements for these conventions, for 
instance, if you use SVG and have several triangle elements belonging to 
different categories according to their appearance, (e.g. color and texture 
and how they move) would it be better to not tell to the user looking the 
markup that they are actually related than use some available mechanisms, 
such as symbols, classes and stylesheets?

And I know very well, there are no conventions for classes but my question 
is should we try to support authors to make them? At least at the site 
level? And how could that then help accessibility? I  would think it helps 
at least offering the users some ready made alternative style sheets. And 
if nothing else it helps managing the site stylesheets.


At 05:27 PM 3/12/2001 +0100, Kynn Bartlett wrote:
>At 11:06 AM -0500 3/12/01, Marja-Riitta Koivunen wrote:
>>These are some things I had in my mind when doing this:
>>- also a nonvisual user knows that the image and the caption text are related
>How?  There is no obvious relationship encoded in the markup.
>>- the caption is not just another paragraph
>>- user can ask to see caption and alt text together if images are not 
>In what user agents can you make this distinction?
>>- user can use classes in his stylesheet to change the type/ color etc. 
>>of the caption so that it is easier to find search for them when visually 
>>browsing through the text or use the class to browse through them by 
>>using braille or speech
>Yes, but using classes in user-defined stylesheets is highly
>problematic, because classes are not meant to be universal over the
>range of web sites.  A user relying upon specific class names can
>easily -introduce- accessibility errors.
>(And of course anything that requires the user to be able to code CSS
>is a bad solution.)
>>- user who want's to read the text first can skip the image and it's 
>>caption and then come back to it later
>>- if user gathers images to a list he can also gather their captions
>How?  What user agents do this?
>Marja, I don't understand how you think the stated benefits actually
>occur from the type of markup you used.  Yes, those are the right -ideas-,
>but saying that encoding the page thusly in (X)HTML produces those
>benefits is incorrect, since I don't believe that any of them actually
>-work- as you'd like them to work.  You could say the same thing about
>encoding it in XML as well, and it probably makes more sense to do
>so since the way you want to use classes is not the way classes work,
>either in practice or in theory.  They are the wrong mechanism for
>doing what you want to do here.
>((X)HTML lacks such a mechanism, which is the root of the problem.)
>Kynn Bartlett <kynn@idyllmtn.com>
Received on Monday, 12 March 2001 13:57:33 UTC

This archive was generated by hypermail 2.3.1 : Tuesday, 16 January 2018 15:33:36 UTC