The quality of Hubble's images could have been improved by introduction of new data e.g. characteristics of distortions, from processing multiple frames etc. It wasn't equivalent of those tricks done in cameras, that only apply cosmetic changes and don't introduce new information.
Modern cameras (including many recent powershots) *do* know the characteristics of the lens (for example, the amount of barrel distortion at a given zoom level), and correct for it. It is directly equivalent. Some resolution is lost when applying this kind of transformation, but again, the resulting image appears to a human observer as a more accurate representation of the original scene.
RAW is the purest *data* you can get. Without processing, it is not the highest quality *image* by any reasonable standard. As the cameras get more sophisticated processing on board, this will only increase.
Do you believe that an image with white balance applied is lower quality than one without ? Hot pixel removal ? De-bayering ?
I really shouldn't continue this argument, but IMO it is important for CHDK users to understand this. There is a strong tendency to expect raws to produce better *images*, users need to be clear raw is just purer *data*. Producing good images is entirely up to the users, and with RAW substantial post processing will be required to even reach the same level of visual quality as camera jpegs.
RAW should not be treated as a final product, it just allows the user to replace the cameras processing pipeline with their own.