I'm not sure, if this was already mentioned in another way, but the main difference between film and digital sensors is the sensitivity to light rays which are hitting the film/sensor at an oblique angle, different from 90 degrees to the film/sensor plane. Film is relatively tolerant to this kind of rays which mainly appear in the corners, far from the middle of the picture. That's why even such compact kameras as the GR1 were able to use 35mm format. A digital sensor reacts very critical to this kind of rays. Because of limited time I only found this article, where you can find a few sentences about "pixel vignetting":
http://www.the-digital-picture.com/Canon-Lenses/Canon-Lens-Vignetting.aspx
The difference between SLR cameras and compact cameras is the distance between the last lens element and the film/sensor. Especially with wide angle lenses, the light is hitting the film/sensor very flat in the corners and the smaller the distance from the last lens element to the sensor/film, the bigger the difference from a 90 degrees angle.
This was also one of the problems, designing the Leica M8, where the distance from the lens to the sensor is smaller than in DSLR cameras. So they put micro lenses onto the sensor surface to change the direction of the light right before hitting the sensor, but even with that kind of technology, they equipped the new M-lenses with a code that tells the camera to balance vignetting by software, according to the lens, attached. I don't know if micro lensing was used in the Epson RD-1, but as you can see in the following review, the vignetting problem is really significant with rangefinder cameras.
http://www.luminous-landscape.com/reviews/lenses/rd-1-lens.shtml
In the DP1, the lens-sensor distance will probably be even smaller, so I wonder how they will manage the problem. Meanwhile, also some pro DSLR cameras use the micro lens technology.
Edit: I forgot to mention, that in my opinion, software vignetting correction is the wrong way to handle this problem, because this way you are boosting areas which contain much fewer information than the correct exposured areas due to the linear design of sensor units. An interesting article is this one:
http://www.luminous-landscape.com/tutorials/expose-right.shtml