There was an earlier question about s-energy; I think I got the point.
However, in various problem settings I see one is interested maximising pairwise distances between points and, there they say, thus logarithmic energy is minimized. Doesn't log grow with distance? To me this means that actually one is speaking about -log(d) where d is distance, isn't he? Is this a convention of some sort?
When logarithmic potentials are in question, are points always considered to be of like sign? I haven't seen any work on oppositely "charged" logarithmic points or sources, whatever they are called. Is this because of the missing practical applications?
Thank you!