The principle of virtual work says that forces of constraint don't do net work under virtual displacements that are consistent with constraints.
Goldstein says something I don't understand. He says that if sliding friction forces are present then the principle of virtual work fails. But then he proceeds to say that this doesn't really matter because friction is a macroscopic phenomena.
The only way I can interpret this is for the friction forces to be a constraint force. But I thought constraint forces were pretty much always forces whose net effect is known but their exact force exerted is difficult to know. For friction we know its force exerted, so why would you treat it as a constraining force?
I also don't understand why friction being a macroscopic phenomena means it doesn't matter for this. Is it because we considering a system of particles?