- From: Robert Burns <rob@robburns.com>
- Date: Tue, 26 Jun 2007 12:01:52 -0500
- To: Smylers <Smylers@stripey.com>
- Cc: HTML WG <public-html@w3.org>
On Jun 26, 2007, at 11:25 AM, Smylers wrote: > > Robert Burns writes: > >> Perhaps I should provide an alternative example just to illustrate >> what I'm trying to say. > > Thanks -- that's really useful. > >> <object >> data="foo.mpeg" >> alt="My kitten fluffy playing with yarn." >> title="fluffy playing with yarn" >>> Fluffy, still only a few inches tall, is playing with a red ball >> of yarn that has to 3 times her size. She has just fallen on her back >> and it looks like the ball of yarn is crushing her. But she's really >> just having fun. </object> >> >> Do the two character strings look different to you in this example? > > Yes, those are different. > > But it isn't clear to me under what circumstances having both these > alternative representations available is advantageous. I can think of > contexts in which either one of them may be preferable -- for > example if > the text farther down the page makes a reference to the "crushing"/fun > incident, then the longer of the above explanations will be needed in > order for the text to make sense to those who haven't seen the > video[*0], whereas if the video is merely an example of your kitten > doing something and the particular action isn't relevant then the > shorter one would be more appropriate. > > But either the detail is pertinent or it isn't. I'm not sure how > if two > versions available the browser should pick which one to display to the > reader. Or, if the reader is presented with a choice how she'd know > which one to go for. Well, I was adding the @title attribute to push the discussion a bit. As for @alt, it may bet that an aural browser user might want to hear a little snippet like that @alt attribute before deciding what order to consume the embedded elements in. The @title attribute might work in place of that. Again, I'm just trying to get us to collectively address these things. Once more the things I think we should be doing: 1) deprecating (dropping) <img> and <embed> and with them the @longdesc attribute that is included only because these elements do not properly support fallback like all other embedded content elements. Instead of the <embed> element we would promote the <object>, <video>, <audio>, etc embedded content elements. Instead of the <img> element we would promote UA support and author adoption of a new <still> (or <picture> or <pic>, etc.) element that supports fallback content. 2) decide whether the inclusion of @title, @alt, and @longdesc on <img> is something that should be extended to the other embedded content elements. In other words images have a @title attribute an @alt attribute as well as a @longdesc attribute that points to a document fragment servving as a substitute for genuine fallback content that all of the other embedded content elements have (with the exception of the <embed> element which I think we should treat the same as <img>.. 3) Consider whether non-visual browser users have become accustomed to @alt and consider whether it might be needed (optionally or required) for all embedded content elements. Or whether the @title attribute could serve that purpose instead (again leaving <img> alone since it would become deprecated along with <embed>). > [*0] Actually I quibble with your "But she's really just having > fun." > If from the video it _looks_ like she's being crushed then anybody > watching the video will be unaware that it was actually fun -- those > who've read the "alternative" content will be better informed > than the > video-watchers! So that fails as a strict alternative. But this is > largely beside the point: I'm sure we could come up with a similar > example which had no such quibble; my main point above would still > stand. Well perhaps the viewers of the video might be able to tell that the cat was actually having fun. This might not come across with the fallback content description I gave without the last sentence. Perhaps we should require aural browsers to somehow enunciate emoticons :-). Of course this would leave the viewers of the fallback <still> image concerned about the cats fait :-). Take care, Rob
Received on Tuesday, 26 June 2007 17:02:03 UTC