This is an archived snapshot of W3C's public bugzilla bug tracker, decommissioned in April 2019. Please see the home page for more details.
When you look at the HTML spec (https://html.spec.whatwg.org/multipage/embedded-content.html#rules-for-updating-the-text-track-rendering), it talks about updating the display for cues that are "showing". And there, it talks about visual rendering for subtitles and captions and for chapters (as a navigation mechanism), and non-visual exposure for descriptions. Rendering of chapters doesn't really work in the way that the "Rendering" section describes - 6.1 should be restricted to captions and subtitles only. Then there needs to be something to describe how this works for chapters and for descriptions. I'm thinking that this doesn't have to be complicated. For chapters we already have special parsing to extract the chapter titles. We can say that the rendering of these chapter titles is expected part of the rendering of the media element controls for navigation purposes and then leave the details to the browsers. For descriptions we could say that rendering is expected to be done by the browser via speech synthesis, once such a feature is available in browsers, and via exposure to screen readers in the meantime. I'm after feedback for these ideas.
In Blink, this is implemented as part of http://dev.w3.org/html5/webvtt/#dfn-rules-for-updating-the-display-of-webvtt-text-tracks I think the place to put it in the spec is "For each track track in tracks, append to cues all the cues from track's list of cues that have their text track cue active flag set." by ignoring cues that don't have kind captions or subtitles.
(In reply to Philip Jägenstedt from comment #1) > In Blink, this is implemented as part of > http://dev.w3.org/html5/webvtt/#dfn-rules-for-updating-the-display-of-webvtt- > text-tracks > > I think the place to put it in the spec is "For each track track in tracks, > append to cues all the cues from track's list of cues that have their text > track cue active flag set." by ignoring cues that don't have kind captions > or subtitles. I'm sure you're right, but as an aside, that quoted sentence is horrible - it depends entirely on italicisation to make any sense, and even then it's hard to understand! The whole section uses words as algorithmic variable names, and those words are also present in the prose as non-variable names. It makes it very hard to read, and impossible to understand when it is read out loud.
(In reply to Philip Jägenstedt from comment #1) > In Blink, this is implemented as part of > http://dev.w3.org/html5/webvtt/#dfn-rules-for-updating-the-display-of-webvtt- > text-tracks > > I think the place to put it in the spec is "For each track track in tracks, > append to cues all the cues from track's list of cues that have their text > track cue active flag set." by ignoring cues that don't have kind captions > or subtitles. This makes the Rendering section wok for captions and subtitles. We still need a description about what to do with chapters and descriptions and probably even metadata, though nothing is done for them other than exposing cues in VTTCue.