"Used as" implies intent and purpose; usually, this is neither the intent nor purpose, and the capacity of an inductor, or the inductance of a capacitor, is an undesirable (yet unavoidable and inevitable) parasitic property of the component.
There are, in fact, situations where these aspects can be harnessed for benefit, but it's usually a very special-purpose thing. For example, almost no one puts these parameters in their datasheets, and they're subject to change anyway. Customized geometry, with these parameters under control, can be quite handy at times.
A distributed transmission line structure is arguably such an example: a transmission line (terminated into an unmatched impedance) exhibits inductance and capacitance at alternate, periodic frequencies.
I would suggest learning about transmission lines. They are the fundamental component, the only one that can truly be expressed by electromagnetism. They aren't so hard to work with; if you're used to describing things in terms of chronological events, nothing else is simpler!
Pure capacitance and inductance (lumped elements) are nonphysical: an abstract circuit diagram (net list, as SPICE calls it) has no representation of the speed of light, and the entire circuit reacts instantly to any given stimulus. It is a useful fiction, which we use only because it results in simple polynomial equations. We inevitably fall into the trap of forgetting the underlying truth, and believing in them directly instead. At which point we must add hacks to our models (parasitics) to make them useful again, and we jolt ourselves awake just a little from our daydreaming.
In principle, it takes an infinite number of capacitors and inductors to model a real component; fortunately, as engineers, we know when enough is enough, so that it's usually quite reasonable to produce a lumped-element approximation. The challenge is knowing to identify transmission line structures, and how to convert them into RLC equivalents suitable for the circuit.
Tim