High quality anti aliasing with DP and x2 or x4 oversampling would be usable only for mono or duo phonic synths because of the CPU overhead. I'm not sure that many people are interested in mono synths.
Other antialiasing tricks are more specialised (such as self PM, BLEPs, MIPMAPs) but are lighter and can be used in poly synths.
Using 2 or 4 interleaved inputs/outputs seems to me the most natural way to do it without adding signal types.
I will use this design for my first synth ("Collage Psyché Synth")... after this project I'll have a better view on the pros and cons of this approach.