The complete solution

New product. Old idea's mixed together.
2. Make controller starting with the Softube control 1. Add 17 faders (Over the EQ), Scrub wheel, and many buttons. Also make a big and small version! Also make bank of 8X8 rotary/push knobs. Two of them should fit right over the channels if placed over the 16 channels mixer.
3. Make MADI sound card native under Sonar. Hardware controls dedicated Chips on MADI card, Like an SSL MX4 w/ 16 times the power. Make a smaller brother with ADAT, USB, and Toslink. Use a good word clock!. This will be one of the larger costs!
4. Don't get into converters, pre-amps, or any of the "Heavy" hardware. Just a PCI card, controller, and software.
5. Switch on controllers changes from the hardware to native/software to whatever. Needs at least 100 possible layers with easy access to at least 5.
6. The big version of the mixer (96 channel) will not make you money, but that's great marketing and could gain you respect with the big boys if you can land a few real installs. A 16 channel version that allows linking would be the mid range and popular. A 4 channel version would be the entry. The hardware will get people hooked on the software. Make the brain replaceable - Upgrades to the buttons, knobs, and sliders. People want solutions, not products.
7. Touch surface approximations and tweeks of the controls as plug-ins/apps.
8. Let the person with the golden ears make some of the choices, even if the costs go up a bit. Assuming Good AD and DA, the MADI to MADI sound quality needs to be AWSOME, not just transparent. TC electronics, Lexicon, SSL, and many others created great software for there time because they let the ears adjust the software settings.
9. This would allow you to write and establish the standard for DAW & hardware integration while only shipping computer cards and control surfaces.
10. This gets divided into three ventures. 1 Control surface. 2 Digital IO/Mixing. 3 Software. Each of the three area's should be a viable profit center needing to work with each other and what's on the market. Cakewalk as a company has experience in all that's required except there is no BIG BOY solution.
11 Think 1 bit, or Very high sample rates. Like 384+. Pitch corrections, Compressor feedback, Amp simulations, all could be better with real fast feedback. Don't think the problem of 44.1 or 48 times whatever will ever go away :(
12 This may be a few years ahead of it's time, but I would think the cost effectiveness of computer has come far enough to make this good solution. Native will continue to grow, but a solid hardware/software solution with a "sound" gains you the market. What is the real cost of a good computer hardware word clock and mixer? Only you could create a great sounding mixing card that is the most compatible with all existing plug ins! SSL has a "sound" associated with them. and they do things digitally. You do need to decide on what the SONAR sound is. And if you say "transparent" think Grace pre-amps. They are so brutally transparent my best description is "Expose all Weakness" sound that is also like digital is without some help from the programmers. This is NOT the sound you want to be identified with. You want create an identifiable great sound quality others don't create. The "sound" may need to becomes a switch in the mixing engine. The switch may need to by dynamic based on the number of channels in the mix.
 
 
Many users can do great things with native alone, but good EQ's, Compressors, and Effects, cant really eat up available computations power in a hurry. A good digital hardware reverb and real compressor(arguable) is still a required component for commercial mixing. is it not about time for a quality complete, cost effective solution?
 
When the sound card gets out of date. that's what gets replaced, the control surface is what needed to still look nice and work well after 8 years. That will get you the re-sell of the sound card. The quality of the word clock and how that integrates with the software for fast feedback within plug-ins is what will give the end product quality.