Michael Yeaney wrote:
I find it interesting that everyone responding to this thread has failed to mention one very important aspect of any design-for-accessibility debate: Until you actually test it with a target audience/persona (i.e., someone who actually **is** blind), we're all just guessing at the relative importance of the issue at hand.
We're not all just guessing - at least I'm not. My base for saying anything on the issue may be limited to national, and even regional, target-groups, but the amount of guesswork I apply is also (to the best of my ability) limited to an absolute minimum. 1: An alt-text should be short (a few words only) and to the point. Anything in addition to that should be found elsewhere, and whether that "elsewhere" is on the page, on the site, or somewhere else on or off the web, depends on the case at hand. 2: the importance of anything on a web page, depends entirely on the individual visitor's interests at the time of visit.
Keep in mind, that some may hear the page read aloud and think 'Sheesh - enough with the graphics descriptions that keep interrupting the text flow of the page'.
The speed at which some hear a page read aloud can be (what we would call) extreme, so we should indeed not interrupt the flow unnecessarily. That's why it's a good thing to "silence" graphics that play no role in the flow, and deliver something useful as part of the natural text flow when it _does_ play a role.
And yes, I've witnessed this, and it is **very** humbling. We can read the specifications all day long and apply them in a (seemingly) 100% correct manner, and yet still totally ruin the experience for some.
Yes, but the main (original) issue here is about making the 'alt attribute' itself optional _in_ a future specification. Reading, and indeed writing, specifications on how to use or not to use the 'alt attribute', only makes real sense if its existence is _recommended_, IMO.
Test, test, test.....
Testing may run into problems with my point 2 above, as few (if any) can afford to do testing with a wide-spread enough panel to "catch all". We have to make educated guesses - and choices - no matter how well we test and tailor our solutions to any group and their recommendations, so "self-education" and "use of common sense" is (at least) as important as testing. regards Georg -- http://www.gunlaug.no ******************************************************************* List Guidelines: http://webstandardsgroup.org/mail/guidelines.cfm Unsubscribe: http://webstandardsgroup.org/join/unsubscribe.cfm Help: [EMAIL PROTECTED] *******************************************************************