Updating first post since the thread has gotten so muddled...
If automation was able to be preserved and the snow could accept 16 channels of automation, it would allow a snow user to preserve these performances and re-adjust patches and parts after they have been recorded to audio without having to shuffle tracks and automation data in their DAW.
Mockup:
[Blockierte Grafik: http://www.straiusmusic.com/Access_Snow_MIDI_Input_Request.jpg]
Selectable inputs for the snow's 4 parts would make it easy to mute MIDI source tracks and continue assigning new automation parts as you move past 4 patches/layers in a project. Returning to a part that you recorded as audio would be as simple as loading your saved preset, re-assigning the part to your desired MIDI input and your existing automation assigned to that channel would continue to be valid.
This wouldn't increase the capability of the snow. It would still have all the same voice and part limitations, but would make it more flexible in rendering your MIDI source to audio and then returning to tweak if necessary.