25737 2014-05-16 08:11:53 +0000 [WebVTT] file using only nested cues definition is too lax 2014-10-10 09:05:45 +0000 1 1 1 Unclassified TextTracks CG WebVTT unspecified PC All NEW v1 P2 normal --- 1 philipj dave.null philipj silviapfeiffer1 public-texttracks oldest_to_newest 106138 0 philipj 2014-05-16 08:11:53 +0000 http://dev.w3.org/html5/webvtt/#dfn-webvtt-file-using-only-nested-cues > given any two cues cue1 and cue2 with start and end time offsets > (x1, y1) and (x2, y2) respectively, > > either cue1 lies fully within cue2, i.e. x1 ≥ x2 and y1 ≤ y2 > or cue1 fully contains cue2, i.e. x1 ≤ x2 and y1 ≥ y2. The following file would match the definition: WEBVTT 00:00.000 --> 00:01.000 A 00:00.000 --> 00:01.000 B Yet, neither cue is nested in the other, and a tree of chapters cannot be constructed from the file. 106140 1 philipj 2014-05-16 08:24:59 +0000 The definition would also fail for something like this: WEBVTT 00:00.000 --> 00:10.000 Chapter 1 00:00.000 --> 00:05.000 Chapter 1.1 00:05.000 --> 00:10.000 Chapter 1.2 Neither of the two last cues contains the other. 106141 2 philipj 2014-05-16 08:46:59 +0000 I think the definition should be that no two cues can have the same start and end time, and the intersection of any two cues should be equal to one of the cues. Not sure how to phrase it in a way that is easily understood. 106142 3 philipj 2014-05-16 08:47:22 +0000 (In reply to Philip Jägenstedt from comment #2) > I think the definition should be that no two cues can have the same start > and end time, and the intersection of any two cues should be equal to one of > the cues. Or empty, d'oh. 106143 4 silviapfeiffer1 2014-05-16 08:59:45 +0000 (In reply to Philip Jägenstedt from comment #0) > http://dev.w3.org/html5/webvtt/#dfn-webvtt-file-using-only-nested-cues > > > given any two cues cue1 and cue2 with start and end time offsets > > (x1, y1) and (x2, y2) respectively, > > > > either cue1 lies fully within cue2, i.e. x1 ≥ x2 and y1 ≤ y2 > > or cue1 fully contains cue2, i.e. x1 ≤ x2 and y1 ≥ y2. > > The following file would match the definition: > > WEBVTT > > 00:00.000 --> 00:01.000 > A > > 00:00.000 --> 00:01.000 > B > > Yet, neither cue is nested in the other, and a tree of chapters cannot be > constructed from the file. "Nested" doesn't have to mean that one can't be completely contained in the other. I think in this case B would be a subtree of A. I think this should be allowed. The first one to come will be the top chapter, the second one the subchapter and so on. I certainly wouldn't want the second chapter to be dropped because it covers the same time interval as the first one. (In reply to Philip Jägenstedt from comment #1) > The definition would also fail for something like this: > > WEBVTT > > 00:00.000 --> 00:10.000 > Chapter 1 > > 00:00.000 --> 00:05.000 > Chapter 1.1 > > 00:05.000 --> 00:10.000 > Chapter 1.2 > > Neither of the two last cues contains the other. Right, I guess there should be a third case where they are totally independent of each other. :-) 106151 5 philipj 2014-05-16 11:10:51 +0000 (In reply to Silvia Pfeiffer from comment #4) > (In reply to Philip Jägenstedt from comment #0) > > http://dev.w3.org/html5/webvtt/#dfn-webvtt-file-using-only-nested-cues > > > > > given any two cues cue1 and cue2 with start and end time offsets > > > (x1, y1) and (x2, y2) respectively, > > > > > > either cue1 lies fully within cue2, i.e. x1 ≥ x2 and y1 ≤ y2 > > > or cue1 fully contains cue2, i.e. x1 ≤ x2 and y1 ≥ y2. > > > > The following file would match the definition: > > > > WEBVTT > > > > 00:00.000 --> 00:01.000 > > A > > > > 00:00.000 --> 00:01.000 > > B > > > > Yet, neither cue is nested in the other, and a tree of chapters cannot be > > constructed from the file. > > > "Nested" doesn't have to mean that one can't be completely contained in the > other. I think in this case B would be a subtree of A. I think this should > be allowed. The first one to come will be the top chapter, the second one > the subchapter and so on. I certainly wouldn't want the second chapter to be > dropped because it covers the same time interval as the first one. A subchapter with the exact same extent seems odd, when would you deliberately author a chapter hierarchy like that? As for being dropped, the WebVTT spec need only be concerned with validity, not processing, which is defined by HTML: http://www.whatwg.org/specs/web-apps/current-work/multipage/the-video-element.html#rules-for-constructing-the-chapter-tree-from-a-text-track (This split across two specs is a bit weird, but I'm not sure in which spec should take it all.) 107526 6 silviapfeiffer1 2014-06-09 13:40:57 +0000 (In reply to Philip Jägenstedt from comment #5) > > A subchapter with the exact same extent seems odd, when would you > deliberately author a chapter hierarchy like that? I've seen book chapters with only one subchapter. I think it's possible and wouldn't want to forbid it. > http://www.whatwg.org/specs/web-apps/current-work/multipage/the-video- > element.html#rules-for-constructing-the-chapter-tree-from-a-text-track > > (This split across two specs is a bit weird, but I'm not sure in which spec > should take it all.) If we can't find anyone else buying into the same hierarchical chapter structure to WebVTT, we might as well move it all into the WebVTT spec. It's what happened with VTTCue. But I am not sure about that either.