I apologize for apparently misunderstanding the scope of what was being proposed.
If a finite set of unencoded Han characters needs to be displayed correctly using IDSes, then the complexity of the look-up tables depends upon how many characters are in the set. It would probably best be handled at the font level and we shouldn't expect any mainstream support. If any reasonable IDS is expected to be displayed as a Han ideograph, then the project would be vast. It's doable and I'm sure there would be several approaches. It would not be feasible at the font level. I agree that the language in the text cited earlier in this thread should be strengthened to clarify that visible display of the IDCs is expected behavior while enabling higher level protocols to remain conformant if they attempt to display constructs in place of IDSes. But I don't think that the fact that a previously unencoded character has become encoded should forbid any application from making a display substitution based on IDSes.
