Does Mohs scale of mineral hardness always hold? According to Wikipedia's article on the Mohs scale of mineral hardness, materials / minerals with a higher rating cannot be "visibly" scratched by materials with a lower rating.
It goes on to admit that microscopic dislocations on the harder material do emerge though by trying to scratch it.
I always thought that physical damage mainly depends on the force being exerted to the material, however the explanation on Mohs hardness suggests that force and work are negligible because there is no mention on how exactly the scratching is to be performed, at which velocity, for how long, at which temperature etc.
This implies that sliding and rubbing two materials against one another will basically have the same outcome regardless of whether the sliding happens through manual movement or at cosmic scales like satellite movement.
Also, would a material be considered "harder" on the Mohs scale if we were to scratch it near its melting point so that it sort of "self-repairs" scratches when being scratched due to friction-induced melting ?
Sorry if these questions sound silly to physics experts and mineralogists.
 A: The Mohs scale is useful because you don't need any special equipment to do the tests so it can be used in the field to help identify minerals. It is not as precise as tests like the Vickers hardness nor is it intended to be. In practice it's usually obvious which mineral is scratching which.
As it happens there was a recent related question at Glass Hardness and Pressure. You need to use a bit of care about how you define a scratch.
A: As John says, the Mohs criterion is useful because it may be immediately applied.
One may try to rob the two materials with any force but the magnitude of the force really doesn't matter because once the force exceeds a certain threshold, the materials' atoms or molecules start to rearrange. Scratches – whatever is their exact definition – will begin to develop and the force you exerted gets reduced.
The point of the Mohs scale is that when the materials start to get modified – develop scratches – to relieve the external pressure, it's far more likely that the softer material according to this scale is the one that will "surrender" first and get damaged by the scratches. This is no exact law. The harder material may sometimes gets damaged, too. But the quantitative difference between the amount of scratches is huge and very sensitive on the scale.
For every temperature, you should in principle quantify the hardness again, from scratch, to use the word again. So materials will surely get softer near the melting point. If they start to "self-repair" because they're partly liquids, this is certainly a proof of their being less hard, not harder! Liquids would have the lowest rating on this scale. A melted piece of the material may not fit the definition of a "scratch" but it is "at least as bad" as a scratch. A material that is really without scratches has individual atoms or molecules sitting tightly at the prescribed points of the lattice or another structure. Freely moving molecules of a liquid violate this rule.
The extended Mohs scale assigns hardness 1 to all liquids, see

http://www.rockroost.com/Mohs-hardness-scale-tips.shtml

A: Unless stated otherwise, when I say a material has a certain hardeness, it means its hardness when it's nanosmooth and has not yet been scratched. There is no way of defining the hardness of any material in such a way that any material can scratch a material of lower hardness but not a material of higher hardness. Different solids have different physical properties, for example, infinitely malleable, partly malleable, completely brittle and amorphous, completely brittle and crystalline. Completely brittle materials do exist because they will not undergo any amount of deformation under any amount of tension or compression stress all the way up to their compression strength. Any material of a given physical property is harder than another material with the same physical property but at a lower strength as I will define absolute hardness in such a way that for any physical property, the hardness of a material with that physical property varies linearly with its strength. Also, for any given physical property, a material with that physical property can scratch another material with the same property if it's not too many times harder but that number varies with physical property. It's even true that for any physical property, the harder a substance is, the less badly it can scratch itself for a given pressing force and scratching speed. A smooth surface of one substance can't scratch a smooth surface of another substance. For any physical property, when a sharp tip of one substance with that physical property slides along a smooth surface of another substance with that physical property, the substance with the sharp tip has to be more times harder than the smooth substance to not get scratched at all than the smooth substance has to be than the substance with the sharp tip to not get scratched at all and the narrower the angle of the tip, the more times harder the substance with the sharp tip has to be to not get scratched at all. Brittle substances also weaken after they get scratched. Some substances than can undergo the glass transition are harder in the amorphous form and other substances are harder in the crystalline form. However, a brittle amorphous material doesn't weaken as much after getting scratched a given amount as a brittle crystalline material. That's probably because when a brittle crystalline solid gets scratched, it initiates a crack which magnifies any tension applied to the material and then if it's put under too much tension, the tension will be magnified at the tip of the crack to higher than the tensile strength in the direction of the tension and then the crack will propagate through the at the speed of sound in the material causing it to fracture. On the other hand, when a brittle amorphous material gets scratched, it also initiates a crack but it doesn't lose as much strength because it doesn't have any cracks that are as long because cracks have more of a tendency to loop around into a chip. The strength a material has when it has a surface etched nanosmooth is called its theoretical strength.
Source: https://en.wikipedia.org/wiki/Fracture_mechanics
A: To add to the replies above, you must remember that hardness tests (like Vickers or Rockwell) measure the ability of a material to resist permanent (plastic) deformation. So in practice, you take a very hard material, like diamond, and use it to indent a range of other materials. You usually don't want the indenter and indented material to have similar hardnesses at all. 
Also, scratch testing is quite difficult to control (as you've guessed).
