> Burning hydrogen moves upward the same as non-burning hydrogen. Gas in contact with flame does not magically become non-buoyant.

Correct, it burns. Like it did in the Hindenburg. I have no idea what your point is.

> We are not, in fact, discussing "hydrogen cars", so whatever "sealed environment" you picture in them is wholly irrelevant.

This is a counter-argument for your previous points that leaks are irrelevant. My point is simple: No, they're not. You're trying to side-step this point by saying "well, I'm not talking about that" when you did right here:

> H2 under pressure moves outward, and in a containment breach tends to, instead, increase the H2 fraction above the 75% that can sustain a flame until most of the gas has escaped and leapt skyward.

Which is complete nonsense if you look into what's around a containment system. Which is my point -- containment is never not in a confined environment. You talk about some positive airflow system, sure, that's fine. What if the failure affects that? Well now you have to consider the risks all over again.

> involved no exploding hydrogen

This is a pedantic nitpick. In day to day usage, people say explosion to mean "rapid, uncontrolled fire". That does not negate my point -- that was one of the largest hydrogen fires in history.

You sound like you have tunnel vision on the advantages of hydrogen for industrial uses. There's a reason we don't use it. It's dangerous, and engineering solutions to mitigate the danger are prone to failure. If you want more: https://www.airships.net/hydrogen-airship-accidents/