Is the use of term Measurement appropriate in quantum physics It may not sound like a physics question, but I think, it is.
When we measure spin of an electron in a certain direction, we are actually, setting it up in that direction. That does not sound like measurement. Wouldn't alignment be more appropriate term for use? 
Doesn't the probability wave give chances of aligning (or not aligning) the particle in certain direction, given a specific spin direction prior to measurement. That specific direction also by the way, we have to set beforehand, otherwise, we would not know what it is.
Given the things at quantum level are so tiny, do we even have the capability for pure measurement, without actually changing them? Then why the process is called measurement. Can we get away by saying we are measuring the alignment?
When we run a fan, why don't we say - oh! we measured the direction of air flow, and it is in the direction where the blades are facing?
 A: Measurement is a process in which by help of interactions one quantum entangles the quantum system under study and the environment and by rapidly utilizing decoherence cleans up the "off diagonals" of the entanglement into an apparent collapsed state. After this process, it effectively appears as the wave function of the system has been set to a distinct quantum state.
This process, entanglement of quantum system on a macroscopic state is used so often that it deserves a special name, and that is a measurement. It is a very usefull tool in studying quantum mechanics, since it is absolutely the only way and only process to obtain quantum information to macroscopic world. Nothing more. It obeys the laws of qm as such and should not be mystified. And it is a distinct process from classical measurement and therefore some of your analogies do not work.
To entangle a system with another, interactions are required, however weak. These interactions perform unitary evolution to the system, which is the measurement process. Thus the measurement cannot leave the system into a same state by definition. Also, the measurement is irreversible, since it is impossible to undo the unitary evolution from thermal fluctuations etc. which caused the decoherence (kind of like the entropy always increases).
However, interesting part of the system can be in roughly definite state after measurement. It is sound to talk about measuring a spin of a particle and then further utilize this particle somehere.
When measuring we often measure things like energy, momentum, spin etc. The reason for this is that one usually measured via gauge bosons and they conveniently carry such quantities to far away from the system at study.
Since the measurement process alters the apparent wave function to an eigenstate of the measurement operator, the state will no longer be the same when measuring again with any non-commuting operator. But that is not about measurement as such, but just a fact of qm that Sx and Sy do not commute for example. One just used a measurement which involed unitary evolution of wave function with interaction Hamiltonian containing Sx or Sy. And this setting the state is just basic qm algebra and not so much related to the actual measurement.
All in all, there is nothing mystical about measurement. To conclude, everything would probably be easier if it wouldn't be called measurement.
