- From: Maciej Stachowiak <mjs@apple.com>
- Date: Wed, 17 Mar 2010 18:34:31 -0700
- To: Ian Hickson <ian@hixie.ch>
- Cc: HTML Accessibility Task Force <public-html-a11y@w3.org>
On Mar 17, 2010, at 6:25 PM, Ian Hickson wrote: > On Wed, 17 Mar 2010, Maciej Stachowiak wrote: >> On Mar 17, 2010, at 5:51 PM, Ian Hickson wrote: >>> >>> Not just accessible content outside the canvas, but accessible >>> content >>> outside the canvas that represents static content that is generated >>> with the page, and not fetched separate from the page. Most uses of >>> canvas I've seen for graphs, including all those I've written, get >>> the >>> data after the page is loaded, and would have to generate the table >>> just like they generate the graph. >> >> Well even in that case you are saved one line of script to >> conditionally >> clear the canvas fallback. > > Why would it be conditional? Because you don't want to clear the fallback in browsers where canvas is not supported. (You also don't want to do any canvas drawing in that case, but you *do* still want to provide alternate content, even if it is generated dynamically). > >> In addition, the alternate form of the data could be behind a link, >> in >> which case whatever is found on the same page as canvas is purely >> static. > > True, though at this point we're way past the 80% rule IMHO. It's hard for me to tell how common it will be to put a more accessible alternative outside the canvas, where everyone can see it or follow the link to it, as opposed to inside. But I think in the case of an alternative visible to everyone, the nonav attribute would make it slightly easier to get it right, and slightly increase the odds that screen reader users won't be faced with an "upgrade your browser" message followed by actual accessible content. I'm still not sure it is a hugely compelling feature, but it does not seem problematic in the same way as prior proposals (like the adom="" proposal or the <accessible> proposal). So while I do not strongly support it, I also cannot find a reason to object. REgards, Maciej
Received on Thursday, 18 March 2010 01:35:05 UTC