As another example of using the dot product, consider the following illustration. It shows how we can ``decompose'' a vector into components along and perpendicular to a second vector. Thus, we can write the vector as which are the vectors in the direction of and perpendicular to a second vector, .
The first step is to compute . We can do this by using the dot product, but we must be careful. Remember that the dot product includes the length of both vectors. The best approach is to find a unit vector in the direction of the second vector. This vector is simply . So the length of is simply the projection of along the unit vector. Thus,
Now, this is simply the magnitude of the part of that is parallel to . We also need direction. But that's easy since this is the vector parallel to , the direction of the vector must be the direction of . So,
The other vector, the component of that is perpendicular to can be found using the relationship . You may be confused with the different ways the word ``component'' is used throughout the subject of vector calculus. I always think of a vector as a single object. When I want to write that vector down, I write it as being made up of parts called components. The appearance of those components can differ greatly depending on how one breaks the vector up. We could use good ole' and so forth, or we could make reference to another vector.
As an example, let's break the vector into its components parallel to and perpendicular to the vector .
Note that if we add the two components above we get the original vector back.