SSA and dominance

Today there was some puzzling about dominance and the placement of Phi, in one ``example'' discussed on the whiteboard, similar to the one on the slide called Improvement (and the core idea of Phi).

What's on the slide is still correct. Also it's correct (in the example from the whiteboard) that the Phi-function needs not to be added at the node n2 in that situation /caused/ by the assigment in n1. However, if in the second branch in the example, there is no assigment (where in the first branch it is), then there will still be a Phi-function added at the discussed point, int's only that addition is not cause by n1, but by the ``other node'' which does not dominate n2.

I will add something like an explanation to the script-version.