Is the intent for it to be possible to use content and shadow elements as fallback content for content and shadow elements?

The algorithm described in "Rendering Shadow DOM Subtrees" has two holes:

In the case of this shadow:

<shadow>
  <content id="a"></content>
</shadow>

When there is no older shadow tree, the algorithm hits this case:

2.(third bullet).2: Otherwise, render fallback content as any DOM subtree

ie since it does not specify "…but with the same rendering exceptions" this means that any nodes distributed to #a are not rendered. Is this intentional?

The opposite case:

<content id="a">
  <shadow></shadow>
</content>

when there are no nodes distributed to #a likewise specifies rendering <shadow></shadow> as any DOM subtree and not with shadow DOM rendering exceptions.

Likewise for

<content select="…">
  <content select="…"></content>
</content>

Could the spec provide commentary on whether the intent? And if the intent is that it is possible to use content and shadow as fallback content for content and shadow, the algorithm needs to be fixed to support this.