Not understanding this way of proving undecidability of the termination problem

I am reading some slides on Algorithm to understand why termination is an undecidable problem. The slides say the following:

– Assume termination(P) always terminates and returns true iff P always terminates on all input data;

– The following program yields a contradiction

while termination(P) skip;  

But I do not see what kind of contradiction there. Any idea?