Camera Transformation

I am trying to write a basic 3d renderer in Visual Basic (i know its not the language for it! but it allows me to focus on the theory and less on whether a pointer is null or not!) Using a derivation I found for the camera transformation I wrote the following code: '' code Public Sub CameraTransformation(T As Matrix) Dim Trans(1 To 5) As Matrix '' calculating t1 MakeIdentity Trans(1) Trans(1).cell(4, 1) = -EyeX Trans(1).cell(4, 2) = -EyeY Trans(1).cell(4, 3) = -EyeZ '' calculating t2 Dim t2_r As Single Dim t2_stheta As Single Dim t2_ctheta As Single t2_r = Sqr(Square(EyeX - LookAtX) + Square(EyeY - LookAtY) + Square(EyeZ - LookAtZ)) t2_stheta = (EyeY - LookAtY) / (Sqr(Square(EyeX - LookAtX) + Square(EyeY - LookAtY))) t2_ctheta = (EyeX - LookAtX) / (Sqr(Square(EyeX - LookAtX) + Square(EyeY - LookAtY))) MakeIdentity Trans(2) Trans(2).cell(1, 1) = t2_stheta Trans(2).cell(1, 2) = t2_ctheta Trans(2).cell(2, 1) = -t2_ctheta Trans(2).cell(2, 2) = t2_stheta '' calculating t3 Dim t3_r As Single Dim t3_sphi As Single Dim t3_cphi As Single t3_r = t2_r t3_sphi = Sqr(Square(EyeX - LookAtX) + Square(EyeY - LookAtY)) / t3_r t3_cphi = (EyeZ - LookAtZ) / t3_r MakeIdentity Trans(3) Trans(3).cell(2, 2) = -t3_cphi Trans(3).cell(2, 3) = -t3_sphi Trans(3).cell(3, 2) = t3_sphi Trans(3).cell(3, 3) = -t3_cphi '' calculating t4 MakeIdentity Trans(4) Trans(4).cell(1, 1) = -1 MatrixMatrixMult T, Trans(1), Trans(2) MatrixMatrixMult T, T, Trans(3) MatrixMatrixMult T, T, Trans(4) ''''MatrixMatrixMult T(), T(), t5() '' for roll - removed End Sub The problem is in the lines: t2_stheta = (EyeY - LookAtY) / (Sqr(Square(EyeX - LookAtX) + Square(EyeY - LookAtY))) t2_ctheta = (EyeX - LookAtX) / (Sqr(Square(EyeX - LookAtX) + Square(EyeY - LookAtY))) When Eyex, Eyey, Lookatx, lookaty are all zero (looking down the z axis) there is division by zero. Does anyone have any idea''s on how I should fix this? Also, is this a good formula for calculating the transformation matrix. After this I multiply all the points (vectors) by the matrix. Should I be normalizing the vectors? Thanks Ian