The electric potential can be defined as V = - ∫C E⋅dl where we are taking the line integral along C from some convenient reference point O, where we have set V = 0, to the point r we are trying to find the potential at. Of course, C can be any curve, but it's usually the most convenient to take it as a straight line from O to r. Doesn't this mean dl will point in the direction from O to r? If that is the case, then say we're trying to find the potential at some distance from point charge +q and we've set our reference point infinitely far away. We have E = (kq/r^2) r(hat), where r(hat) is the spherical unit vector. Then, E⋅dl = E r(hat)⋅dl. But dl points from infinity to the point we're trying to find, and say that we've picked O so that dl points in the -r(hat) direction. Then, E⋅dl = -Edr, so evaluating the integral, we get V = -kq/r, which is evidently wrong. I am aware that we can fix this by taking O to be somewhere else, but do you always take the line integral to be positive? If so, why? Or is there something else I'm missing? EDIT: This seems like it'd fit better in General Physics. Woops.