I'm creating my own transformation pipeline for fun. Found this information in DX SDK:

http://msdn2.microsoft.com/en-us/library/bb206260(VS.85,printer).aspx

In step 7 they have this:



But where do they get that W from? Is it the W component of the vector after being transformed in step 6? It doesn't have a subindex

X, Y and Z don't have a subindex neither, but I suppose that X = Xm, Y = Ym, Z = Zm, and W = Wm = 1, as they said at the beginning:

If a vertex in the model coordinate is given by Pm = (Xm, Ym, Zm, 1), then the following transformations are applied to compute screen coordinates Ps = (Xs, Ys, Zs, Ws):
OTOH that doesn't have sense, since all the transformation would be a division by a factor from the pipeline. Then got no idea

Thanks!