The issue with the bump not updating sholud be fixed in the latest dailies.
As for the strength, it's a bit complicated. The percentage is just a relative value, which scales the perturbation that would normally result from performing the bump mapping calculation. Perturbation just offsets the surface normal in some direction by a given amount. As an example, if the normal points straight up and the bump mapping perturbation is 1, the resulting angle is 45 degrees. By increasing the strength to 200%, you'll get a perturbation of 2, which would result in an angle of atan(2/1) ~= 63 degrees. 300% will be 72 degrees, etc...
But then again, you don't know the amount of the perturbation to begin with, without knowing the exact algorithm that calculates the bump mapping perturbation. So right now, I'd suggest just trying it out and use the amount, which feels right. We might think of adding some units later ...