Hello, after experimenting with the Virtual Dataset feature, it seems to correspond to our needs, and is completely transparent to the reader. Thus, the specification will not need to change to support these new features. The original post has been edited to reflect that. Thanks again for your feedback!