I don't know what else DSI can do in the analog realm that'll be mind blowing except for multitimbrality...I'd love to see Dave tackle that....not sure he would though. I just feel anything else is redundant at this point when it comes to analog. Probably why I didn't go for an OB6 or Deep Mind 12...it's just a mono timbral analog synth with very little difference to my Prophet 6. I'll probably go for a Rev2 at some point just cause it's different but even still...duo timbrality has been done decades ago...let's move on.
I used to think the same way about multi-timbrality, though it's tough to do well without a more elaborate MIDI master keyboard controller framework (like the Ensoniq stuff had, but that the DSI synths lack). I'd rather see split/layer done well than multitimbral done poorly. (I'd guess that the Tempest falls into the multitimbral category, but lacks the need for keyboard zone mapping.)
I understand that but I'd rather see something all under the hood. I just never understood how it's possible that vintage synths could have more features and capability than newer ones.
I don't think it's about features and capability, per se, but more about usability. The Poly Evolver is a perfect example of that which can go wrong with a multitimbral implementation; the Tempest, on the other hand, works quite well (in its guise as a drum machine).
I don't think we'll see another A6, largely because it was a failure outside of its niche (setting aside its sound-generation technology). And I'd say that the Waldorf Q was equally guilty of the same issues (even though it was a DSP-based, virtual-analogue instrument); both suffered from a lack of a well-designed polyphonic sketchpad sequencer, a feature that would appeal to musicians across a broader range of genres and styles than, say, arpeggiators / step sequencers of a particular opinionated bent.
Barring that sketchpad functionality, the additional parts would probably be underutilized by most players (or could easily be accommodated instead by using external sound sources, pointing back towards my remark about master controller functionality being a fundamental requirement).