How about mechanically superb construction that wouldn't become obsolete the second sensors brakes down, specially after global shutter technology becomes available at affordable prices and there wouldn't be mechanical shutter to brake down.
Let's face it, those two things keep the milage on digital cameras rather low, shutters that brake and sensors that go bust. When talking about seriously well build cameras that is.
1. The alignment and positioning of the image sensor plane are *extremely* critical. Much more so than most realise. Setting up a system where users can routinely remove and replace sensors without screwing up this extremely critical adjustment would be, at best, notrivial.
Such systems would also add mechanical complexity and weight.
2. Sensor storage, cleanliness, and damage would be major issues.
3. The hardware portion of imaging pipelines is for the foreseeable future progressing faster than sensor tech. A new sensor with an old imaging pipeline would be stupid. Maintaining interface compatibility with the camera's computational systems (storage, AF, UI, exposure, etc.) would be limiting (since these systems would not be modular, presumably) and, again, would impose major design and manufacturing penalties over time.
4. At the prices that would likely have to be charged for such a modular system, the market would be small, further increasing costs due to a lack of economies of scale.
IMO, modular sensors would create many more serious problems than they would solve.
You watch. If someone is dumb enough to introduce a system like this in anything but high-end medium format, they will take a bath.
And if the modular system is (as rumored) from Olympus, it will be the third major system camera that they abandon in less than two decades (OM,
4/3, modular). I do love Olympus optics (especially their microscopes), but this is an error that they cannot afford to make.
Canon could afford to make such an error, but they are not stupid enough to do so.