"Power" includes economy of expression. If 5 lines of language A requires 500 lines in language B, language A is more powerful.
"Power" includes things like directly accessing hardware.
Deconvolution is a tricky game. For one thing, as you mentioned, the deconvolution matrix varies across the frame, and for another, it often must be much larger than the related convolution matrix (PSF). It is often ill-conditioned (effectively, that means there's a number close to zero in the denominator), and it always magnifies noise. If there are saturated regions in the image, they can't be handled properly.
Deconvolution and other postprocessing can improve images, but turning a grey mess into a beautiful full color photo is something that can usually be done only under tightly controlled conditions. All too often, there's something screwy in the results.
It takes a mirror to be a reflex camera, and it is only related to the shutter in that the mirror has to not block the light to the sensor/film when a picture is taken. Twin Lens Reflex (TLR) cameras use a fixed mirror, unrelated to the shutter. Beam splitter Single Lens Reflex (SLR) cameras like the Canon Pellix and Canon EOS RT use a fixed mirror unrelated to the shutter. Conventional SLRs use a flip-up mirror that moves before the shutter opens and flops down after the shutter closes.
Reflex refers to the optical path to the viewfinder, involving a reflection.