- From: Kynn Bartlett <kynn-edapta@idyllmtn.com>
- Date: Mon, 12 Mar 2001 16:32:06 +0100
- To: Marja-Riitta Koivunen <marja@w3.org>, w3c-wai-gl@w3.org
At 9:31 AM -0500 3/12/01, Marja-Riitta Koivunen wrote: >I often use the definition below for images and their captions. I >though it enhances accessibility and makes it easier to change >styles. I can see how it would help with styles. But I'm not sure exactly how it would help with accessibility. >So should we say something more explicit or is it that there are no >accessibility benefits in doing this and I should stop? No reason to stop, necessarily. ><div class="figure" id="Fig-1"> ><p><img src="architecture.png" >alt="Basic architecture showing the communication between the client >and the annotation servers."></p> > ><p class="caption">Figure 1: The basic architecture of Annotea.</p> ></div> It's somewhat interesting but I'm not sure what you actually do with the information you're encoding here. How you style and/or transform this content would create the value here, not just the encoding of meaning/intent. (Current browsers don't, to the best of my knowledge, do anything useful with this.) What you're trying to capture here would be expressed in (arbitrary) XML as something like: <figure id="Fig-1"> <image src="architecture.png"> Basic architecture showing the communication between the client and the annotation servers. </image> <caption> The basic architecture of Annotea. </caption> </figure> ...but then how you use that information will be very dependent upon how you use it. ("Style and/or transform" in my paragraph above.) (X)HTML unfortunately doesn't have a rich enough set of presentation to express what we want, but there are several ways to render this which might work -- such as using the <table> tag to group related content: <table summary="Figure Fig-1: The basic architecture of Annotea." class="figure"> <caption> Figure 1: The basic architecture of Annotea. </caption> <tr> <td> <img src="architecture.png" alt="Basic architecture showing the communication between the client and annotation servers."/> </td> </tr> </table> (The above can be generated from the XML using XSLT relatively easily.) Or, as you've done in the original, you can use <div> for this kind of thing. Or some other presentation depending on what you're trying to accomplish. Note that this example is by necessity limited; no longdesc is given which fully explains the architecture, and the caption and the alt attribute provide pretty much identical information -- a more complete example would be something like: <figure id="Fig-1"> <image src="architecture.png"> Basic architecture showing the communication between the client and the annotation servers. </image> <caption> The basic architecture of Annotea. </caption> <longdesc> <p> When a client requests an annotation from a blah blah blah blah blah blah </p> <p> ... and that's how it works. </p> </longdesc> </figure> Or possibly just: <figure id="Fig-1"> <image src="architecture.png"> Basic architecture showing the communication between the client and the annotation servers. </image> <caption> The basic architecture of Annotea. </caption> <longdesc href="arch_desc#1.html"/> </figure> My basic comment is that you wouldn't want to write your _content_ in the manner you describe -- you hopefully would want to write it in non-XHTML XML -- but that the way you produced output markup from that content is okay but not the only way to present that type of information. (Which serves to illustrate a weakness with (X)HTML -- multiple ways of presenting the same semantic mean that it is a poor language for encoding semantics!) --Kynn -- Kynn Bartlett <kynn@idyllmtn.com> http://www.kynn.com/
Received on Monday, 12 March 2001 10:40:21 UTC