The reason it's so hard to do a digital M is that the light strikes the film plane at a steeper angle than an SLR, and sensors can't cope with this very well, film of course doesn't mind. The assertion is that this same problem crops up on cameras like the the 1DsMkII at the corner of the frame and is frequently mis-identified as a problem with the lens.
That is an old urban legend, dating back prior to the introduction if the 1Ds. There is a rather easy solution to this "problem", which involves laying the microlenses down on the sensor surface with a slightly narrower spacing than the sensor pixels. The microlenses are centered over the pixels at the center of the sensor, but are offset inward at the edges, so that light striking the sensor at an angle can be efficiently utilized. In any event, failure to use this technique would only increase vignetting, but would have very little effect on chromatic aberration. Given the fact that most 1Ds-film comparisons show similar levels of vignetting when the same lens is used on the film and digital bodies, there's no real-world evidence that this theory has any validity. If the lens is closer to the sensor, all that is necessary is to increase the edge microlens offset accordingly.
The real issue is that silicon records more image data than the same area of film, making lens flaws that used to be hidden in the grain structure visible and obvious.