I was looking at the sourcecode for a camera class where the view matrix was created a little differently to how I was used to, probably because It doesnt use any createlookat type helper functions.

In this code the look vector, up vector, and right vector were calculated for the view which I understand.

but then the dotProducts between the position of the camera and each of the vectors was calculated and used in the fourth row of each column.

Does anybody know of a resource that explains the maths behind this step? I understand whats happening but I dont understand why.

Also the elements in the fourth column are all set to 0 accept the last one which is 1.0 does anybody know what this is used for?


Thanks so much.


EDIT: In case its relevant the code uses Directx