Why a zero dot product still trips people up
Many courses stress that orthogonal means dot equals zero, then move on.
The uncomfortable detail is the zero vector: if either arrow has length zero, the angle story collapses, cosine is not meaningful, and you should read the scalar as a raw sum of products instead of trusting a geometric slogan.
This workspace keeps both stories visible. You always see the numeric sum. The angle chips grey out the cosine when a length hits zero so you do not pretend θ still exists.
From typed components to one number
The tool multiplies each matching pair of coordinates, adds those products, and shows the running layout in the open “Term breakdown” drawer.
Parallel to the sum, it computes the Euclidean lengths of A and B, derives θ through the standard arccos guard rail on the ratio (A · B) / (|A||B|), and prints cosine when both lengths stay positive.
Projection rows appear only when both lengths are positive. They reuse the same dot value you already see in the headline scalar.
Same job, three different notations
Authors switch symbols depending on the textbook. The table below lists one plain-language gloss for each form so you recognize the same operation under a new name.
| Symbol habit | What you type here | What the page returns |
|---|---|---|
| A · B with coordinate tuples | Matching rows x, y, z, … | Sum of aibi plus angle readouts when lengths allow |
| |A||B| cos θ | Same tuples (θ inferred) | θ and cos θ chips after lengths are computed |
| Row vector times column vector | Still the same coordinates | Identical scalar, because matrix multiply reduces to the same pairwise products |
When you already keep everything inside three dimensions and want cross products or subtraction in the same session, the vector calculator page bundles those operations next to the scalar lane.
Where this page refuses to help
Floating-point arithmetic means cosines near ±1 sometimes drift by a tiny epsilon after rounding, so arccos input is clamped to [-1, 1] before the angle call.
The plot is a teaching aid, not a CAD viewport. It uses a fixed isometric-style projection for 3D and a standard xy plane for 2D, so lengths in the drawing are not measured in physical units.
If you need symbolic simplification, exact radicals, or variable letters instead of decimals, use a computer algebra system. This interface stays numeric.
Quick scenes that use A · B
- Physics homework on work. With force (3, 0, −2) newtons and displacement (1, 4, 0) meters, the scalar work preview is the dot of those tuples before you attach units in your write-up.
- Graphics lighting checks. A surface normal and light direction both become triples. Their dot controls how strongly a Lambert shader brightens a pixel before you normalize colors in the engine.
- Text embeddings (simplified). Two short score vectors from a model behave like high-dimensional arrows. Cosine similarity is the dot divided by both lengths, which is why length chips matter before you trust an “similarity percentage.”
For pure perpendicular magnitude in ℝ³, pair this scalar check with the cross product calculator when you need a vector orthogonal to both inputs.
