The ratio of two forces acting respectively perpendicular and normal to a contact surface of two bodies, the coefficient of friction, is widely used in engineering and science depicting the friction resistance of materials sliding over one another. Ruled by the so-called Amontons-Coulomb friction laws (independence from the load, the contact area and the sliding speed), this dimensionless quantity appears to be convenient for engineering and relatively easy to determine. Nevertheless, the use of tabulated friction coefficients becomes somewhat an issue to predict friction behavior of mechanical systems. The system dependence of friction is sometimes ignored, leading to misapplication. Moreover, the fundamental origins of sliding resistance are not as clear and care should be taken when attributing a fundamental significance to the friction coefficient. This paper aims to clarify findings on friction Charles Augustin Coulomb did and that have been used for hundred of years. At the demand of the French Académie Royale des Sciences calling for rules and accurate data to design machines, Coulomb developed an experimental study of friction of woods and metals used for rubbing parts in machinery. Alongside his findings, he was aware of limitations of Amontons' findings. He demonstrated by experimentation the dependence of friction on the normal load, on the area of contact or on the sliding speed. In his day, he proposed some physical interpretations in agreement with his findings. He explained friction was originating from a variety of sources which led to various friction behaviors.