I've been advised off-list that my attempt to make an analogy with CJK doesn't sit well.
It's fair to say that ideographic variation sequences are for plain-text representation of material which isn't suitable for atomic encoding. An analogy can be drawn from that situation to the situation of other scripts, such as Latin (or Khmer).
The ideographic variation sequences also represent an anomaly: if it's not suitable for plain-text encoding, it doesn't *need* plain-text representation. Except that it does.
It's the demands of the CJK user community which drive the plain-text representation, which is proper. This method should apply to non-CJK scripts as well.
Styled Latin text is being simulated with math alphanumerics now, which means that data is being interchanged and archived. That's the user demand illustrated.
Whether the users are doing it Chicago style or just plain willy-nilly doesn't matter; it's being done. User communities drive their own script development and advancement using the tools available.
