Thanks Simone for writing so clearly about the topic. I think eventually, photography textbooks should be rewritten to better reflect current technology and the physics of imaging. With film, sensitivity was largely fixed and most film types were available in only one or two speeds, and the photon detection wasn't as efficient as it is today. Photographers would use film and they'd know that fast film yields grainier images, but most didn't really stop to think why it is so. It's because recording only a small amount of light leads to uncertainty in the measured quantities because of the randomness of the arrival times of individual photons. This is the not the complete picture (there are other sources of noise which fortunately tend to become less significant as technology matures) but it's the one factor which cannot be removed by any advance in technology, thus it is important to understand. Today we have sensors where the ISO can be adjusted across a huge range and the image quality (in terms of SNR) closely approximates the limits of what is theoretically possible. This has meant that the imaging systems can be now understood using a theoretical framework which considers relatively simple physical principles first, and leaves the complex implementation details out (at least in the first approximation), yet this theoretical analysis gives useful results which are close to what is experienced in real life using these new systems. Hence the models can be helpful to aid understanding how different cameras can be used to create similar results (certain things are not dependent on the imaging system but stem from the fundamentals of the physics of imaging and optics) and also to elucidate the underlying physics of why results using given settings and setups are different.
I can understand that a photographer working in practice needs to use the tools that are in hand and cannot replace them with tools that might exist only in theory. However, it cannot hurt to try to understand basic principles in different ways. Richard Feynman was famous for his ability to come up with different analogous models to explain physical and mathematical phenomena. This kind of thinking is very useful and deepens one's understanding. Of course, none of these explanations is really the "why" of how things work the way they work in nature. That kind of question cannot be answered. What we know is certain basic principles such as the conservation of momentum can be used to explain and predict a lot of physical phenomena and these principles are widely used in engineering. It should be the same with photography. The usefulness of a model or theory is validated by applying the model or theory to predict the result of a new experiment.
Today many sensors are close enough to the "ideal sensor" that simple principles can be applied with great success, as shown by the dpreview.com article on equivalence. There is no "snake oil" being sold here. Equivalence predicts, for example, that if the camera position, shutter speed, angle of view, and depth of field are set the same, then the signal-to-noise in the final image will be about the same as well in the final print (of the same size, viewed from the same position), independently of the sensor size. This is a very important result and an important validation of the usefulness of the theory. It is helpful because it tells us as photographers that at diffraction limited apertures, noise is really a function of how much depth of field you want, not a question of which format you choose to work with. I'm assuming here that subject movement (e.g. due to wind or another reason) forces us to use a certain shutter speed and thus shutter speed is not a parameter we can play with. Thus a macro photographer who needs reasonable depth of field can work with whatever format they find the most practical and will not be punished in terms of image quality until the lens becomes aberration limited instead of diffraction limited. Another, interesting result from theory is that if we set depth of field so that the lens is diffraction limited, at equal depth of field, all formats suffer the same amount of blur due to diffraction. Essentially small formats are very good tools for photographing small subjects whenever you need a lot of depth of field. Larger formats can give the same results but they are larger and more expensive. However, what favours larger formats are the aberrations at wide apertures. For shallow depth of field work, larger formats are preferable because relative to image size the aberrations play a smaller role (at least in central areas of the image) and you get to choose from a larger range of apertures which give "good enough" detail in the final image. All of these results can be understood by theory and are confirmed by practical experience.
However, in certain cases an individual tool has a distinct advantage. For example there is no DX equivalent to the 24/1.4 on FX, a lens that would give a similar picture on DX as that lens at its maximum aperture gives on FX. Another thing is that if you are not comparing equal shutter speed cases, then you may find that one tool works better. For example there is no DX equivalent to a D810 set to ISO 64. Not in terms of resolution of the final image nor in the dynamic range. But these are "fringe" areas where the implementation of a particular camera is unique and offers an advantage. Similarly a DX camera can give more detail in good light than an FX camera using the longest lens you have. Thus not everything can be replaced by an equivalent something else.
