What is the locus of points $(s,t)$ equidistant from the convex side of parabola $y = x^2$?
Now, it is quite easy to determine s and t separately as functions of x, but can you define t as a function of s? Can it even be done?
When the question is applied to the concave side, something interesting happens.