Amazingly, only one reader has emailed me to point out a problem with this image, released by ESA on November 20. I didn't notice the problem either. Look carefully.
Images don't come down from spacecraft in familiar formats like JPEG or TIFF; each instrument produces a data stream that is unique, different from previous instruments. Sometimes instrument teams write their own software to display images, but sometimes, especially if they are in a hurry, they may throw their data into some easier-to-use, off-the-shelf software package that doesn't necessarily know such details as the orientation of the columns of the CCD in the spacecraft camera. Ultimately, this particular mistake happened because, in the middle of a critical spacecraft maneuver, scientists were trying to squeeze in some observations of the planet that would be interesting to the public, and get them out there to the public as fast as they could -- an admirable goal that will occasionally result in a mistake or two. This one is easy to correct, anyway; here's my version of the Rosetta image of Earth, flipped and also rotated so that north is approximately up.
I don't want to hear anyone being a Monday morning quarterback, criticizing Rosetta or ESA for the error! The way to avoid making such mistakes is to take more time for review; but, in this case, ESA's fast release of images allowed the public to ride along with the excitement of a critical spacecraft maneuver, and I think that was important. Presumably the OSIRIS team will have the kinks worked out of their procedures by the time they actually get to their mission targets -- asteroids and especially comet Churyumov-Gerasimenko.