Great to see you become so enthousiastic so quickly.
However: As happens so often in these hallowed halls a simple question starts to move in all sorts of hardly expected directions as soon as one posts it. So I do fear that it is already time to drag you back to the concepetual phase first, even if some of you might only go there screaming and lkicking.
I do not have anything against digital input devices, such as rotary encoders (with or without led rims) but please consider this:
If we look at the Axoloti in isolation we only have 5 available digital in/outputs. Working with 5 rotary encoders will quickly lead to the old mistake of trying to see the whole house through a letterbox. Every time you want to get aquiainted with any of the houses occupants or objects they will first have to be placed in front of the, all too tiny letterbox. This is actually what made most early digital synths difficult to stomage. The Yamaha DX 7, which promissed to be able to do everyhting but was a pig to master, immediately springs to mind. Believe me, I've been there.
5 inputs are very meagre in that respect anyway. Such stuff only tends to work well if you add a display directly above the controllers reminding you of the present address of the inputs. But that display would need a digital output from the Axoliti itself, bringing the number of digital input controllers back to 4. Then you'd need one controller to actually select the submenu, bringing the number of menu data input channels back to 3. Etc.. See where I am going?
Yep, I see 2 counter arguments coming anyway:
1 - MIDI can provide loads of control numbers and even work with data dumps. True, but the world is full of nifty and affordable MIDI control devices so why not buy one of them anyway.
OK, the best interface ever will never be marketed. So if that is your thing just go ahead an build a wall sized personal controller anyway. Been there as well.
I pesonally however have a clear reason to not go to MIDI. I am totally hooked on using polyphonic aftertouch and stuff. So MIDI already has enough trouble anaging pure musical input data when I am around. That alone will clutter up your MIDI data stream quickly enough and lead to all sorts of MIDI lag situations if you want to include a lot of addional traffic. So in that respect the direct in and outs of the Axoloti board are not only a bonus but can actually be a lifesaver.
2 - 5 rotary encoders could of course be used to work in conjunction with the analog inputs. So why not hook up all controls you want "permanent" acces to in your instrument to the analog inputs and add a small rotary coder matrix as an additional source for more esoteric parameters.. The rotary encoders could then provide an almost endless number of tiny submenus.
The disadvantage of that would however again be that one will loose oversight very quyickly. You'd again need lots of notes, overlay and spreadsheets to keep your mind from exploding. Thast is actually why I propased a "simple" 5 x 15 matrix. Anyhting more then that will be difficult to keep track of for mere mortals anyway.
But even if you would personally be able to wrap your own mind around an even bigger matrix: Will the Axoloti itself then not become cluttered up with all the data management routines needed for such a (theoretically) limitless system?
As a samller expansion on my original pushbutton selction idea this could however still be viable
Wrapping things up: I do have no qualms with using digital input dervices but I'd prefer it if we, at least for now, concentrate on ergonimically viable direct Axoloti hardware controllers, which are as we have already established limited to 5 digital ports anyway. In my opinion that makes an analog / digital input hybrid unavoidable.