[whatwg] Alt attribute for <video> and <audio>

[Ian Hickson]

On Fri, 14 Aug 2009, Silvia Pfeiffer wrote:
> On Fri, Aug 14, 2009 at 9:13 PM, Ian Hickson<ian at hixie.ch> wrote:
> > On Mon, 10 Aug 2009, Remco wrote:
> >>
> >> Shouldn't <video>s and <audio>s (and maybe <object>s too?) also have 
> >> an alt attribute? A quick Google search tells me this has not been 
> >> discussed before.
> >
> > For users who can use audio but not video, authors should either 
> > provide audio descriptions in the video file as alternative tracks, or 
> > supplemental material provided in links available to everyone near the 
> > video.
> >
> > For users who can use video but not audio, authors should provide 
> > subtitles, captions, or transcripts either in the video or audio file 
> > as supplemental tracks, or in supplemental materials available to 
> > everyone in links near the video.
> >
> > For users who can use neither video nor audio, supplemental materials 
> > are likely the best thing for an author to provide, again, in links 
> > visible to all.
> >
> > For users of legacy UAs that don't support these features, 
> > feature-rich alternatives such as plugins can be provided as fallback 
> > content for <video> and <audio>.
> >
> > Captions and subtitles can be included either directly in the media 
> > file, or scripts can manually support external resources using the cue 
> > range API. Going forward, we will probably also support dedicated 
> > formats that UAs can merge with the video to handle showing external 
> > subtitles natively.
> >
> > I don't see a need for an alt="" attribute here. What problem would it 
> > solve that is not solved by the above solutions?
> 
> There is only one thing I can think about that an "alt" attribute could 
> provide that nothing else does: as a blind user tabs onto a video 
> element, the "alt" attribute's content could be read out and briefly 
> describe what is visible in the poster image - or alternatively give a 
> brief summary of the video. This is useful for all those cases where no 
> surrounding text is given for whatever reason. Where a surrounding text 
> is given, such as the video title and description, such text is likely 
> not necessary.

It seems that ARIA attributes, the title="" attribute, <figure><legend>, 
and just regular UA defaults (e.g. announcing that you're on a video 
element) are sufficient.


On Fri, 14 Aug 2009, Silvia Pfeiffer wrote:
> On Fri, Aug 14, 2009 at 11:09 PM, Henri Sivonen<hsivonen at iki.fi> wrote:
> >
> > I believe aria-label addresses this.
> 
> Excellent. Then I haven't seen a good argument to add it. Let's not.

On Fri, 14 Aug 2009, Remco wrote:
> 
> Yes, I think that covers it. This also covers the most important, but 
> apparently always ignored case: authors who don't have time for 
> accessibility. A significant portion of web authors will not provide 
> subtitles for every published video. Nor will they provide links to a 
> transcript. Even if they care about accessibility, it's just not 
> economically viable to do it. The best you can hope for is a sentence or 
> two explaining what the video does.
> 
> This also covers other non-text elements: <iframe>, <embed>, <object>.
> 
> The only thing left is ARIA's integration with HTML. Have you had 
> success with your draft? http://hsivonen.iki.fi/aria-html5/ I see you 
> only had one reply to your first announcement. Will the remaining ARIA 
> attributes be an explicit part of HTML? Will the "aria-" prefix be 
> removed?

ARIA is now integrated in HTML5.

-- 
Ian Hickson               U+1047E                )\._.,--....,'``.    fL
http://ln.hixie.ch/       U+263A                /,   _.. \   _\  ;`._ ,.
Things that are impossible just take longer.   `._.-(,_..'--(,_..'`-.;.'