My question here is derivative of my last one but for increased brevity of both questions I've separated them. I'm exploring the following function for learning:$$f(x,\ y) = (x-2)^2 + (y-2)^2$$
And a graphic (courtesy of GeoGebra):

Now, for my question. How would I graph $f(x,\ y) = x^2 + y^2$ by hand? Would I have to assume $x$ and $y$ both as inputs, and then evaluate them for $z$? I surely can't approach it the way I do with a function of $x$ only, where I can evaluate everything at different $x$.
I'm leaning towards it being extremely hard or near impossible to graph them hence why we use computers, but knowing exactly why is what I'm most keen to understand. Since there are two inputs, any arbitrary pair of alike or unalike inputs of $x$ and $y$ respectively have an output $z$, i.e. there is a $z$ for $x = 2$ with $y = 0.22222$ just as there is a $z$ for $x = 1$ with $y = 1$, and thus too many points to plot. Is this the correct thinking? If not, why, and what is the correct thinking?
How do we graph normally? In the beginning we pick an x, get out a y. Then as you get better at graphing certain graphs become very familiar and you don't have to pick a ton of x values. For example $x^2+y^2=1$ you can recognize that graph is a circle with practice and without picking points. For 3D it's a similar process. Start by holding a variable constant. When $x=0$ we have $z=y^2$ which is a parabola. This translates to "along the plane $x=0$ we have a parabola". Repeating this process for both x and y we realize that we have all these parabolas and we can recognize our graph as a paraboloid.