This is an archived snapshot of W3C's public bugzilla bug tracker, decommissioned in April 2019. Please see the home page for more details.
The current text defines the mapping of track information to track language and label twice per media resource format: once for the texttracks and once for the audio/video tracks. Most of the time it is the same. This is error-prone. I suggest defining it once, unless it is different.
kind is always different. id, label and language are often the same, though in MPEG-2 label seems to be different depending on kind. While I can see your point, I find it easier to read it this way. But if you can make it work, I'm not against it.