forgot to mention that pixel size is a trade off between High Dynamic Range and Resolution. A future digital camera will utilizes technologies employed by Nokia's 1020 and Samsung ISOCELL. Nokia delivers better signal to noise ratio due to averaging mulitple pixel data, while ISOCELL allows for better dynamic range.
So soon, you will get cameras that can maintain high resolution (avoid aliasing) and perform well in dynamic range.
The real revolution to come in imaging is when they get beyond digital graphics and start using vector graphics (used in the likes of Corel Draw and other high end programs), If they develop vector graphic cameras then the lines are defined by a mathematical formula, which, produces the same image whether at 1 cm X 1 cm or 1 m X 1 m.
Nevertheless, cameras will not be consigned to the ash heap of history, because the optics involved in producing the image remain.
Digital graphics works by dividing an area into little sub areas (pixels) and reproducing an image by mapping the image and trying to get it duplicated in the pixels. The limitation is you can only be so accurate because of the limitation of the number of pixels. Hence the use of anti-aliasing logarithms in both computer display and photography. While there were a few others who championed vector graphics Corel was the major developer of describing shapes by mathematical vectors. This approach gives an "accurate" reproduction by using a mathematical vector rather than mapping pixels. If a curved line is adequately described by Y = X squared -3X + 2 it is mapped mathematically and pixilation or noise can be reduced to the limit of the digital reproduction. However, the beauty of this approach is that you can (with enough detail) create an exact analog reproduction. The mathematical formulae are infinitely more accurate than the digital maps. The problem is setting a common standard for describing the vectors and reproducing them. With contemporary computer processing power the use of vector graphics not only becomes possible but preferable to digital reproduction. You can spend all day tracing an S curve digitally and not get the accuracy or detail of the mathematical equivalent.
I like the infographic, but I think how big you can print depends on more than just the megapixel count. In fact, I'd argue that other aspects are more important, like a having a sharp and well exposed photo with no noise to begin with. See <a href="http/www.disturbancesinthewash.net/journal/how-many-megapixels-do-i-need">this article</a> for example.