The top of a ladder is clearly a step. It’s inscribed with “this is not a step” only to avoid litigation. The disclaimer is totally unrelated to the stepness or non-stepness of the top of the ladder.
(But please everyone: follow all manufacturer safety instructions.)

Also because we always have a shoulder in the tone curve, a white diffuser will alway be proportionally closer to middle gray in the final rendered image than it was in real life. (I mean someone COULD make a literal/linear rendering but that would look very electronic and not artful.)

So: in the final rendered image as opposed to the real life scene, a white diffuser can have any luminance at all (depending on the preferred LUT, grading style, etc) whereas scene white is fixed at 10x middle gray.

Diffuse white (paper white) is not the same as my scene white. Because although a white diffuser has a fixed relationship to middle gray in the real-world scene, in the photographic rendering, there’s no fixed relationship: authors can choose any style they like: gentle shoulder, sizzly whites, etc.

This is a very compressed format with few words allowed in messages. I was merely replaying to the topic you addressed to me, not trying to be rude. I just didn’t use the limited characters for pleasantries. Apologies.

(I literally say this in the checklist section that the wasted space and coarsening of the steps is the same with chroma as it is with luma: that the wider gamut uses up space in the container and makes the steps coarser for any given bit depth.)

Yup. And I discuss all this in the demo -- it's not contrary to what I said in the demo. Maybe you missed part.

I never said that the number of increments is linearly proportional to the bit depth: in fact, I explicitly say that 8-bit as 256 steps and 10-bit has 1024. That is, as you say a doubling for each bit. But it is true that the portion of those steps that are above scene white in HDR is about a third.

Not sure why you're saying "ouch" and complaining about me wasting your time when you're the one who came at me saying that I got something wrong. I'm not required to agree with you and I'm explaining why I don't. (I'm open to being wrong but don't think I am in this case).

You’re still saying that something’s “not right” but you’re not identifying what that is. If something’s not right, I’ll release an updated version. So far, the only thing you’ve identified is the “linear” thing, which isn’t in the demo so no update is required for that not to be in the demo.

I never said that compression is “linear.” I only said that it’s a real world problem that actually (not just hypothetically) degrades images, so can’t be ignored. So, this is a straw man complaint that’s unrelated to the demo.

Isn’t your job the one where you tell the computer what CGI to make and then it does it?