RIGHT, Let us talk about phsical laws. A small sensor will only need a shorter focal length to give the same angle for a bigger sensor. Micro 3/4 has a mutiplicayion factor of 2. So the 14 mm Panasonic ancake lens will be equilvalent to a 28mm lens on a full frame. The 14mm will need a total of 40mm (20mm for the lense, 20mm for the fringe depth). That is a ration of 2.86. Now apply this ration for the 28mm for FF. than it will need 80mm. That is physical Law. Sensor size does not matter??? The bigger the sensor, the bigger the lense for the same viewing angle.
Thr Fringe diatance can be set to anything by the manufacturer to suit the usage. NEX fringe is set so thin for the using of adapter for the OLD range finder lenses,like Leica, Canon etc. The fringe distance can be set independent of the back focal plane distance. For mirrorless, the lenses have the luxury of going inside of the body and into the body. Have you look at the Leica wide angle lenses lately??? As the the rumores 1mm back focal distance lens from Canon, it is going to be huge. In order for it to work, the rear element of the lens MUST cover the whole sensor. What it mean is the lense will ge going inside of the fringe and way into the body. You just cannot make a cmera body with a 1mm fringe distance.
Here comes the phsical law again. the size of the optic is determined by the focal length, speed of the lens and the associated mechanical and electronics part. It got nothing to do wit hthe fringe distance (except for wide angle lens for range finder cameras, that is another story). Just look at the pancake lens from Panasonics. The optics are small, the lens dismeter is large. That is for the mechanical and theelectronis component. It has got nothing to do with the fringe distance.
You've got your physics a little mixed up and the topic/areas being discussed here too it seems.
When we talk about back focus plane distances and how it affects the size of the lens, we are referring to sensors of the same and equal size, not across different sizes. When you get one design with a back focus plane at 1mm versus one at 30mm, the optics required for the shorter one is smaller. Par equal area of focus and within limits of optics diffraction. Which is why Leica comes into such a picture. It was design to use FF (in the film days) but yet for equivalent focal length and aperture lens, they were smaller than say Canon or Nikon designs, even if they were all full manual, non-AF lenses.
In this case of Canon's mirrorless, nobody seems to be certain for sure what's the sensor size. And we can only infer from the back focus plane distance that for the sensor size they are designing for, they are trying very hard to miniaturize the size of the lenses. And for what ever size sensor they are using. Be it a FF, 1.3x, 1.6x or 1.85x crop, it SHOULD be (within limitations of cost and practicality) smaller than lenses designed for sensors of equivalent crop factor.
You are the one that mixed up the sensor size. I started out with the Canon 40mm pancake lens which is supposed to be for the FF. You switched the discussion to micro 4/3 and use it to justify a smaller lens.
My physics is right. The reason why some Leica lenses are smaller than DSLR lenses is that the 35mm lens for Leica does not need the retro-telephoto design that is manditory for DLSR. When you look at the 21 mm Leica lens (it is a retro-telephoto design). It is not small at all. This proves that the fringe distance does not decrease the size of the lens. Even the 50mm Summicron, it is longer than the 50mm Canon DSLR lens. The longer Leica lense is just as big as the DSLR lenses. FYI, f stop is define as the focal length/diameter of the aperture. This will limit the minimum total size of the lens for a given focal length, regardless of the fringe. I suggest you look at the Leica lenses physically before any further discussion. just as I said before if there is a 1mm back focus plane lens exist, the rear element must cover the ENTIRE sensor to make it work. Therefore this lens cannot be small. Did you take a class in optics???
Well you could do it, but the challenges in keeping the aberrations in check would be daunting, to say the least.
On the other hand, if the sensor itself were sufficiently curved, you could get as close to it with a normal rear element as you please. A combination of the two approaches could be reasonable, while significantly slimming the total package.
Seriously though if anyone has it right it's Intel. They don't break what's working, they just slim it down a little more on every other iteration, using the ones in between to improve the current process. It works because no single step is too ambitious, so they always make forward progress and never have complete flops. Keep doing that for a few years and you can have some seriously killer products.