And it wouldn't have been much easier to program, because it used a novel form of sound generation which most people still find hopelessly unintuitive.
A Germany company made a programming box for the DX7 with one-knob-per-feature. It was very expensive and sold very badly, even among professionals who could easily afford one - which suggests that the market for programmability really wasn't there, even with knobs.
People don't care about knobs. What they care about is simplicity. Synths like the Microbrute require virtually no mental effort. The sound design model is so simple virtually anyone of average intelligence can learn it in an afternoon. Within a few days they'll know most of what there is to know - which is how to make the same few stock sounds that analog synthesizers have been making since the 1970s.
This is fine for beginners but experts find it very limiting. Brian Eno loves his DX7, and the fact that you can set up the envelopes to make quasi-generative noises is a big part of the appeal. Most people don't get that far with it, but those that do aren't going to be happy with the very obvious limitations of a Microbrute.
Touch screens in cars are more complex. Having a map in a car is a useful thing, and there's no muscle memory. It's hard to see how you can replace drag/touch to set a destination with buttons that are less annoying and more intuitive.
But IME the big problem with touch screens is that they don't work. The screen in my car is resistive, so I have to keep poking at it to get a touch to register. The hybrid design - two rows of soft buttons, one knob, one screen - would be fine if the touch part actually worked.