Everything from the 4 lessons of this topic on a few pages: the key idea of each lesson, the diagram to remember, the points to learn and the traps examiners set. Read it, cover it, test yourself.
A vector in space is \(x\mathbf{i} + y\mathbf{j} + z\mathbf{k} = \begin{pmatrix}x\\y\\z\end{pmatrix}\), with magnitude \(\sqrt{x^{2}+y^{2}+z^{2}}\). The displacement from \(A\) to \(B\) is \(\overrightarrow{AB} = \mathbf{b} - \mathbf{a}\): end minus start.
\(\overrightarrow{AB}\) is \(\mathbf{b} - \mathbf{a}\), not \(\mathbf{a} - \mathbf{b}\). Getting it backwards gives \(\overrightarrow{BA}\): the same length but the opposite direction, which ruins every later step that uses the direction.
Parallel vectors need ONE common factor for every component. \((2, 4, 6)\) and \((1, 2, 4)\) look alike, but the factors are \(2, 2, 1.5\), so they are not parallel.
The midpoint is half the SUM of the position vectors, \(\tfrac{1}{2}(\mathbf{a} + \mathbf{b})\). Half the difference, \(\tfrac{1}{2}(\mathbf{b} - \mathbf{a})\), is the step from \(A\) to the midpoint, not the midpoint itself.
The line through the point with position vector \(\mathbf{a}\), in the direction of \(\mathbf{b}\), is \(\mathbf{r} = \mathbf{a} + t\mathbf{b}\). Each value of the parameter \(t\) gives one point on the line, and every point on the line comes from exactly one \(t\).
Finding \(t\) from ONE component proves nothing. A point lies on the line only if the SAME \(t\) works in all three components; always check the remaining two.
The direction is \(\mathbf{b} - \mathbf{a}\), never the position vector of the second point. Writing \(\mathbf{r} = \mathbf{a} + t\mathbf{b}\) with \(\mathbf{b} = \overrightarrow{OB}\) gives a line through \(A\) parallel to \(OB\), which usually misses \(B\) completely.
The parameter is not a distance unless \(\mathbf{b}\) is a unit vector. With \(\mathbf{b} = (4, 2, -2)\), \(t = 5\) means \(5\sqrt{24}\) units, not 5 units.
First compare the directions: if one is a multiple of the other, the lines are parallel (or the same line). If not, set the two general points equal, solve two component equations for the two parameters, and test the third: it holds when the lines intersect and fails when they are skew.
In two dimensions, non-parallel lines must meet, so students skip the check. In three dimensions non-parallel is not enough: you must show the lines share a point before calling them intersecting.
Using the same letter for both parameters (say \(t\) and \(t\)) forces the two points to have equal parameters, which they almost never do. You then wrongly conclude 'skew'. Always use \(s\) and \(t\), or \(\lambda\) and \(\mu\).
Intersecting PATHS do not mean a collision. The parameters are separate; a collision needs both objects at the point at the same time, which means equal times, not just equal positions.
\(\mathbf{a} \cdot \mathbf{b} = |\mathbf{a}|\,|\mathbf{b}|\cos\theta = a_{1}b_{1} + a_{2}b_{2} + a_{3}b_{3}\). So \(\cos\theta = \dfrac{\mathbf{a} \cdot \mathbf{b}}{|\mathbf{a}|\,|\mathbf{b}|}\), and non-zero vectors are perpendicular exactly when \(\mathbf{a} \cdot \mathbf{b} = 0\).
The angle in the formula is between vectors pointing away from the same point. For the angle \(ABC\) in a triangle use \(\overrightarrow{BA}\) and \(\overrightarrow{BC}\); using \(\overrightarrow{AB}\) and \(\overrightarrow{BC}\) gives \(180^\circ\) minus the angle you want.
Use the DIRECTION vectors of the lines, not the position vectors of the points on them. The position vectors depend on where the origin is; the angle between the lines does not.
Dot \(\overrightarrow{PN}\) with the DIRECTION of the line, not with the position vector of \(N\) or of \(P\). And remember to finish: the question usually wants \(N\) or \(PN\), not just the parameter.