Foundations — Distance from a point to a line

This page earns every single mark of the parent formula before the formula is allowed to appear. The one symbol we can name right now, because it is the goal, is $d$:

Everything else — $a$, $b$, $c$, $x_1$, $y_1$, the bars, the square root — is built below, in the order that lets the next symbol make sense. Nothing is used before it is built. Only at the very end, once every mark has meaning, do we assemble the formula.

1. The plane and its two number-arrows

Picture it: look at the figure. Right is positive $x$, left is negative $x$; up is positive $y$, down is negative $y$. Every location on the sheet gets a two-number address.

Why the topic needs it: the whole problem — a point, a line, a distance — lives on this sheet. Without addresses we cannot say where anything is.

2. A point $P(x_1,y_1)$

The little "${}_1$" in $x_1$ is just a name tag. It does not mean "multiply" or "power one". We write $P(x_1,y_1)$ to say "this particular point, whose horizontal address I'll call $x_1$ and whose vertical address I'll call $y_1$." If a second point turned up we'd call it $(x_2,y_2)$ — same idea, different tag.

Why the topic needs it: $(x_1,y_1)$ is the field-you're-standing-in. The formula's whole job is to measure this point's distance $d$ to a line.

3. What a line equation $ax+by+c=0$ really says

How to picture "makes it true": pick any dot on the drawn line, read off its $(x,y)$, plug into $ax+by+c$. You get exactly $0$. Pick a dot off the line and you get some non-zero number — positive on one side, negative on the other. The line is the razor-thin boundary where the expression flips sign.

Why "$=0$" form matters: the parent formula reads $a$, $b$, $c$ straight off this exact layout. If your line arrives as $y=mx+k$, you must first shove everything to one side — $mx-y+k=0$ — to see $a=m$, $b=-1$, $c=k$. (Prerequisite: Equation of a Line.)

4. The letters $a$, $b$, $c$ — and why $(a,b)$ points across the line

Here is the quietly magical fact. The pair $(a,b)$, read as an arrow starting at the origin, points perpendicular (at a right angle) to the line. This arrow is called the normal vector.

Why it's perpendicular (in plain words): move along the line from one point to another. The equation's value never changes — it stays $0$. The direction in which the value changes fastest is exactly the arrow $(a,b)$, and "fastest change" is always at right angles to "no change". So $(a,b)$ crosses the line square-on. (Deeper: Normal Vector to a Line, Perpendicular Lines.)

Why the topic needs it: the shortest walk to a line is along this perpendicular arrow. The formula measures $d$ in the direction of $(a,b)$ — no other direction gives the shortest path.

What the constant $c$ does: it slides the line

Keep $a$ and $b$ fixed and change only $c$. The direction of the line never changes (the normal $(a,b)$ is untouched), but the whole line slides bodily across the plane, parallel to itself.

5. An arrow with length — the vector, and $|\vec{n}|$

To find the length of the arrow $(a,b)$ we use the right triangle it makes with the axes: it goes $a$ across and $b$ up, so by Pythagoras its length is $|\vec{n}|=\sqrt{a^2+b^2}.$

Why we must divide by it — units. The score $a{x}_1+b{y}_1+c$ is not a length yet. To see this, imagine doubling $a$, $b$ and $c$ together: the equation $2ax+2by+2c=0$ describes the exact same line, yet the score at any fixed point doubles. A genuine distance cannot depend on how we happened to write the equation. The cure: the score always scales in step with $|\vec{n}|=\sqrt{a^2+b^2}$ (double $a,b$ and $|\vec{n}|$ doubles too), so dividing the score by $|\vec{n}|$ cancels that arbitrary scaling and leaves a number that only depends on the line and the point — a true length in coordinate units.

6. The absolute value $|\cdots |$

Picture it: distance on a ruler from $0$ — always to the right, never negative.

Why the topic needs it: the score $a{x}_1+b{y}_1+c$ is negative on one side of the line and positive on the other. But a walking distance $d$ can never be negative. The bars $|\cdots |$ strip the sign, keeping only "how far".

7. Assembling the formula — and two ways to trust it

Now every mark is earned. Take the score of the point $(x_1,y_1)$, strip its sign, divide by the length of the normal:

Two independent derivations (both fully worked in the parent note) confirm it. Here is the why of each, in one breath, using only symbols we now own.

Route A — the projection route (needs a unit normal)

Pick any point $Q$ on the line and draw the arrow from $Q$ to your point $P$. Its shadow cast along the unit normal $\hat{n}$ (its projection) is exactly the perpendicular gap $d$ — because $\hat{n}$ points straight across the line. Working that shadow out algebraically, the "$Q$ on the line" fact $a{Q}_x+b{Q}_y+c=0$ makes the $Q$-terms vanish and leaves precisely $\frac{|a{x}_1+b{y}_1+c|}{\sqrt{a^2+b^2}}$.

Route B — the area route

Take two points $A,B$ on the line and your point $P$; they form a triangle. Its area can be found two ways: by the coordinate formula (Area of Triangle Using Coordinates), or as $\frac{1}{2}\times (\text{base }AB)\times (\text{height }d)$, where the height is the perpendicular distance. Setting the two expressions equal and cancelling the base $AB$ leaves the same $d$. (Both routes rest on the Distance Formula, and minimising such distances is the seed of Least Squares Regression.)

The must-check case: $P$ on the line gives $d=0$

If the point already sits on the line, then by definition its score is $0$, so $|a{x}_1+b{y}_1+c|=0$ and the whole fraction is $0$. No walk needed — you're already there. This is the single most important sanity check: the algebra agrees with the picture at the boundary.

Prerequisite map

Equipment checklist

Test yourself — each line hides its answer. Say it out loud first.