I do agree ur words. Once upon a time, when I was a student, the materials on text books seemed not so difficult to understand.
but after working in industry, I found there exist so many non-ideal, non-linear, noisy elements which were once ignored in the text books. the easy formula became more complex and more strange. and someone who was senior in analog design would tell u that "in experience" ...... and seldom "in theory".......
This is really bad, and it always make things more hard to control and the learning cycle become longer in time.
But "from my own expericence", the theory is always the best friend whenever I faced strange problems. If u want to "really understand" ADC, u should design one ADC with the real spec. from the "real world company", not from paper.
and I suggest u can design ur ADC first from behavior models, Matlab is a good choice. After u know the operations of ur ADC in behavior, u can modified ur model to more elaborate and accurate with some non-ideal properties, and do more simulation with different conditions. Then u should know what u are going to design in transistors level, and u will have confidence to design ur ADC.
that's my experience, share with u.