OK, I will play the devils's advocate, and suggest that the Leica Digilux 2 (aka the Pansonic something-or-other) might at least be a succesful "concept camera" for a certain photographic niche. Like a concept car, it will not necessarily sell in large quantities though, or at all.
Apart from the appeal to some of us of a compact digital camera that allows one to control focus, aperture and focal length with rotating wheels on the lens barrel, the big debate will be over sensor and pixel size, and the EVF. I will leave the latter to another post.
The key here is noting that smaller formats can typically be used at lower f-stop's, and hence at lower ISO settings, off-setting their weakness when compared at equal ISO, and that bigger sensors can only provide their potentially greater image quality when provided with enough light to avoid underexposure.
Larger sensors can give higher image quality when provided with enough light, but an "information theoretic" approach suggests that with a given, limited amount of light, any sensor big enough to take in all the available light without highlight blow-outs, combined with a fast enough lens, can give "image detail" as good as a larger sensor, which has to be unerexposed with that same light levels. Below a certain moderate level of total available light, even 2/3" format can roughly match any larger format. However the lens speed of the Digilux 2 is not enough to fully take advantage of this: I dream of a future f/1 zoom lens in this very small format.
My measure of "detail" is roughly resolution (pixel count) times tonal detail at each location (S/N ratio). If the same total amount of light is gathered (same exposure duration, same aperture diameter and hence f-stop in proportion to focal length and sensor size), sensors of different sizes with the same number of photosites will each get the same amount of light per site, but the larger sites will have more thermal noise, so worse S/N ratio. If the larger sensor instead has more photosites of the same size as the smaller sensor, each photosite gets less light, so again worse S/N ratio, balanced against higher resolution. Downsampling to the resolution of the smaller sensor will return one roughly to the situation of both sensors having the same photosite count; the smaller one still wins on noise.
For the Digilux 2 used at its maximum of f/2 with a shutter speed of 1/100, the sensor can handle all available light at about EV 9 or below (6 stops or more below bright sunlight), rising to EV 10 if 1/200 is needed and so on, so for a larger sensor to have an advantage in light this low or with shutter speeds at least this fast needed, it can only do it by going to larger aperture diameters and hence less DOF; as far as DOF, the equivalent f-stop is about f/5.6 with APS DSLR's. Allow an APS DSLR a zoom lens going down to f/2.8, and the threshold where the Digilux 2 is at no disadvantage goes down to EV 7 at 1/100, EV 8 at 1/200 and so on. Optimal f-stop for image quality seems to be smaller; around f/11 in APS format, and using that and the Digilux 2 at the equivalent f/4 pushes the break-even to about EV 11, just four stops below bright sunlight.
As to the "concept camera" idea, the potential is that (a) the maximum usable light level per pixel might be increased by several stops by ideal like latest Fuji's SuperCCD design, which would increase by several stops the light level needed before a larger format has any advantage ( in the very small 2/3" format, a f/1 lens should be no harder to make than f/4 for comparable FOV in 35mm format, and that would add two stops of high shutter speed potential © the resolution can be increased; Sony has already announced 8MP in 2/3" format, and that is more resolution than the overwhelming majority of prints need (being significantly more than given by any APS format DSLR!)
Finally, do not be mislead by the rather common claim that more, smaller pixels on a sensor of the same size must give worse noise in an image: smaller pixels in the same quantity on a smaller sensor do that, but increasing pixel count in proportion to pixel size roughly balances out the noise level seen in a print of a given size, and also balances any apparent loss in fineness of tonal gradations and range, all through the extra "dithering" or "averaging" effect of printing more, smaller pixels. A print can have far wider dynamic range than the individual pixels: consider half tone printing where the many tiny printed pixels are pure black or white, or consider the pieces of black metalic silver on a traditional print as "1-bit pixels".
In summary, I think that there is a reason why sensor makers are moving almost uniformly towards more, smaller pixels despite the horror often expressed on the internet at this trend, and I think this is the explanation.