I'm just getting familiar with analytical geometry and basically I have two probably very simple questions.
Question 1:
Let's assume I have some plane, i.e: $\pi: x - 2y + 4z - 8 = 0$
What should I do to find points on that plane?
My take on this: to find any point $P(x_{0}, y_{0}, z_{0}) \in \pi $ I can assume that two out of three $x_{0}, y_{0}, z_{0}$ are any constant values $\in \mathbb{R}$ and then find the value of third coordinate. For example let's assume: $x_{0} = 0, z_{0} = 8$, then:
$$0 - 2y_{0} + 32 - 8 = 0$$
$$y_{0} = 12$$
$$\Rightarrow P(0,12,8) \in \pi$$
Is that how should it be done?
Question 2:
Let's assume I have some line i.e.: $\displaystyle \frac{x}{2} = \frac{y - 1}{-2} = z$
I can directly see that point $P(0, 1, 0)$ will surely belong to that line. But let's say I want to find any other point(s) - how should I do it?
I suspect that I can switch the line equation into the parametric form, and then plug in for example value $5$ for parameter $t \in \mathbb{R}$.
Is this the way to go?
Your first idea is precisely what I would do (and have done before): on a plane, one has $2$ degrees of freedom of movement, corresponding to the fact that one has freedom in choosing values for any two of the variables$^\dagger$, and this will determine the third. Your second idea is also what I would do, wherein the line can be parametrized as:
$$\alpha(t) = \begin{bmatrix} 2t \\ -2t + 1 \\ t \end{bmatrix}$$
The trick, for anyone feeling like I pulled this out of a hat, is to simply let any one of the variables be the parametrizing variable $t$. Here, of course, we let $z=t$, and then, using the given relations, solved for the other variables so that we obtained expressions $x = f(t)$ and $y = g(t)$; these functions $f$ and $g$ serve as the $x$ and $y$ components of $\alpha(t)$ respectively. Finally, note that parametrizations of curves are never unique, so don't panic if a purported parametrization and yours don't perfectly visually compare.
P.S. if you have two points $P_1$ and $P_2$ on any line in $\mathbb{R}^n$, one can parametrize this line as $\alpha(t) = (1-t)P_1 + tP_2$. Indeed, check that $\alpha(0) = P_1$ and $\alpha(1) = P_2$, and you can see $\alpha$ is linear because, for instance, it has constant derivative. Parametrizing in this way has added utility: one can tell where points are relative to $P_1$ and $P_2$. Note that points corresponding to $t \leq 0$ lie "before" $P_1$; points corresponding to $0 < t < 1$ lie between $P_1$ and $P_2$, and so forth.
I only mention this last bit as this discussion reminded me of a coding project I'm working on where this idea saved me so much time (needed to implement a subroutine that checks where a line intersects another relative to two points on the line being intersected). Hopefully in my sharing, it serves someone equally well.
$^\dagger$This "...any two of the variables..." bit isn't quite true in every case. For example, if the plane was horizontal with $z=1$ fixed (or, in general, when the plane is parallel to one of the planes spanned by any two of the coordinate axes), clearly one can't choose an arbitrary value for $z$ (or whichever variable is fixed). But one nevertheless still has two degrees of choice freedom: $x$ and $y$ can be anything.