Mountpoints are an inherent design part of NormalizedNode API, as it expresses a special case in terms of a ForeignDataNode.

As such, mountpoints form a separate tree, similar to metadata. This needs to be expressed in NormalizedNodeResult class hierarchy.

As such, MountPointNode needs to be renamed, as it is not a NormalizedNode, but rather:

• we have NormalizedData, which does not imply PathArgument
• we have NormalizedNode, which implies PathArgument and is a tree
• we have NormalizedMountData, which is attached to a NormalizedNodeResult

Note that NormalizedMountData has an identifier, and that identifier has a QName, but it is not a NodeIdentifier. This implies we need to separate NormalizedNodeContainer from NormalizedNode – as the latter implies PathArgument. This will also pave the way for dealing with rc:data, which is NormalizedNodeContainer, but its identifier is a raw string (and therefore is not a NormalizedNode, but a top-level construct as the one introduced in this issue).