If I understand correctly you are having problems in understanding how we arrive from 1.4 to 1.5. So we consider in 1.4 that projection of x along w and consider that number to be greater than some number (bias: b). We can think of b-1 as another number c, so the choice of number 1 is more like one unit distance. If you follow the derivation our optimization rule in 1.10 will still be valid for some other constant. Because the constant term (bias: b) eventually cancels out. Does that clear your confusion? Thanks for visiting.