I've found that with hardware design, at least at the beginner level, you can read up and cram all the theory you can get into your brain, and not really be able to build much. I think that by actually doing, and having someone show you the basics of "this is the way we do something, and this is why we do that something" is much more helpful, and opens up realms of possibility.
I remember from learning to program that there was a point where, probably because of good teaching, I "got it" and then I was able to teach myself. But I haven't yet gotten to the autodidact point in electronics. I've been looking for the book that gives me this "got it" experience!
In my EE work, I tend to use formal analogies in my thought processes, sometimes unconsciously. For example, water and electricity behave in same way [1]. You don't design the whole circuit in the water domain, but you might temporarily flip over to water to understand a particular aspect, then flip back to electrical. Typically I won't write anything down in the alternative domain, using it just enough to clarify my thinking in the electrical domain, then recording things in electrical terms. There's all sorts of other domains too: acoustics, mechanics, pneumatics, ... Pick the one you are most comfortable with in the situation.
In short: Learn constant coefficient 2nd order differential equations, and you can design a whole slew of mechanical, electrical, and hydraulic contraptions.
A real application of this: before computers, mechanical systems control engineers would write the equations of their mechanical systems, write their controllers, and then implement their controllers around the electrical circuit that duplicated the behavior of their mechanical system. Sort of an analog computer way of simulating a system.
Or as in my experience as an EE, learn how to use Laplace transforms, find a table for the 2nd order differential equations, use Laplace transforms, forget how to solve differential equations.
