The logic in 10.6.2 for resolving height when height is auto can be a bit confusing, mainly due to the ordering of conditions and the repeated use of “Otherwise.”

The section says:

If height and width both have computed values of auto and the element also has an intrinsic height, then that intrinsic height is the used value of height.
Otherwise, if height has a computed value of auto, and the element has an intrinsic ratio, then the used value of height is:

(used width) / (intrinsic ratio)

Otherwise, if height has a computed value of auto, and the element has an intrinsic height, then that intrinsic height is the used value of height.

The way the first “Otherwise” is written makes it seem like the intrinsic ratio is only used when intrinsic height isn’t available which isn’t actually the case. I suggest removing the word “Otherwise” and simply stating:

If height has a computed value of auto, and the element has an intrinsic ratio, then the used value of height is:

(used width) / (intrinsic ratio)

It’s also confusing that intrinsic height appears twice once when both dimensions are auto, and again later when only height is auto. This raises the question of what happens when both an intrinsic height and an intrinsic ratio are available, especially when width is defined. The spec doesn’t explain how to resolve that situation or which one takes precedence.

Consider revising the last “Otherwise” to:

If height has a computed value of auto, and the element has an intrinsic height and not an aspect ratio, then that intrinsic height is the used value of height.

There’s also a case that isn’t addressed at all: when both height and width are auto, and the element only has an intrinsic ratio (with no intrinsic height or width). According to 10.3.2, that means the width is undefined and if width is undefined, then calculating height from (used width) / ratio doesn’t work either. This seems like an intentional omission (because width is undefined, so height can’t be resolved), but it’s not obvious unless the reader is familiar with how 10.3.2 interacts with 10.6.2.

I suggest adding a short note to 10.6.2 to make this edge case clear. For example:

Note: If both height and width are auto, and the element has only an intrinsic ratio (with no intrinsic width or height), then width is undefined (see 10.3.2), so height can't be resolved from the ratio either.

It’s a small clarification, but it would make the relationship between these two sections easier to follow.