As I wrote after equation 1.4 that, we don’t know anything about the number b (known as bias), it is better to not consider the effect of sign at this point. You can just think of it as a scalar b = -c. I just did not want to use many notations to confuse readers.

For equation 1.5 it won’t be -1 at this point. But as I have mentioned, once we want to merge the equations for known samples on left and right of the decision boundary, then we can introduce a new variable y_i, which is +1 for samples on the right and -1 samples on the left. Hope this helps. Thanks for your comment. Don’t forget to check the references given in the text. They will help you clarify the doubts.