No! Don't do that -- that's not a great idea, sorry to say. ;)

I'm sure you know Bresenham's algorithm, no? (If not, just post that you don't and I'll discuss this in more depth!)

Basically, you can just use Bresenham's algorithm, and the error variable divided by the larger of the two distances (be it X distance or Y distance) is your percent alpha. So, for the current pixel, you'd use cur_error / max_error for the alpha, and you would also plot the pixel at the current location plus the step (that is, if X is greater you'd move along the Y) with an alpha of 1 - cur_error / max_error .

I hope that made some sense ...