This will not end well
You’re making more than the usual number of mistakes and, given Hixie’s imperviousness to any evidence he doesn’t agree with, your spec will without question end up being so flawed it cannot and will not be used.
Don't we already have the Timed Text spec for this sort of thing?
Contrary to Hixie’s claim, I didn’t make any “mistakes”
My edits to this spec are based on 30 years’ experience in captioning, which is 30 years more than Hixie has. He knows next to nothing about the topic and his edits, which he glossed as “mistakes” (18:52, 21 April 2010), obliterate known use cases – known, that is, to actual captioning viewers.
I say again: This will not end well. It’s another example of Hixie’s petty tyranny. The only upside is that petty is the only tyranny he can manage.