If the reader can't see the semantic markup, then what's the main purpose, and how does this purpose justify the additional effort required, both to implement this system in a standardized way across journals, and by authors who have to tag their citations with this semantic markup?
It's for the algorithms that process papers to produce metrics. Display systems can choose to ignore it and show them all as [1] or might use color, hover message, or whatever to show the distinction in a way that doesn't alter the standard reading experience if that's desired.
