Hi Til,
Is one way of doing it, which I am thinking of doing too. Although I am just puzzled why Karl G designed it this way. I don't want to jump to that route too fast, we may be missing something that Karl G has thought of.
So far what I can think of are:
1. so when we get the relative path on CR we will not end up with /content_folder/content_item/content_symlink. That will make content_folder irrelevant since content_item can be in a non leaf path.
2. content_symlink name might get duplicates since its not a child of content_folder which enforces a constraint that name must be unique.
So far that is all the good reason I can think of why Karl G is wanted to make content folder as the parent of content_symlink. I am basing those from the code of content_symlink__new.
If no one else posts then I guess I can rest my mind.