The difficulty is that the rules are usually rather vasgue about what constitutes manipulation. It can't be banned entirely, because it's long been considered OK to do the kinds of general overall manipulation which correspond to old fashioned darkroom work, such as adjusting exposure, choosing contrast, bringing up shadow detail, changing black & white cutoff points, sharpening, etc.. Then there's the selective manipulation which correspond to the old darkroom dodging & burning, choosing different parts of the image to lighten, darken, etc, as landscape photographers often do to bring in cloud detail along with shadow detail.
There's a documentary aspect to press photography which helps. Press & documentary photographs are meant to represent reality, to show what a person being there at the time, possibly looking through binoculars, might have seen. Asking that the photographer submit the RAW file for checking the extent of digital manipulation rules out all those photographs which come from more than one RAW file, such as replacing a boring sky with one taken at a more interesting time. It also rules out the use of multiple photographs taken at the same time, such as with two different exposures, one for the sky detail, one for the ground, as Ansel Adams sometimes did, or stitching a number of panned images together to get a panorama of a cityscape or a huge crowd at a political demonstration. So one might want to make an allowance to permit the use of multiple exposures, i.e. multiple RAW files provided they were all taken from the same position and within one minute of one another.
But taking two photographs from the same position within one minute of each other is also the way some levitation photographs are made, where a person seems to float above the ground because the two images make it easy to remove the person's support. That can be taken care of by adding the rule that no pixels are to be changed which add or subtract something not present or absent at the time of the photograph.
That raises the question of what constitutes a thing seen in the scene. Obviously one should not remove lamp posts or add people. But is a shadow a thing which should not be removed? What about light sources? Shadows tell us how many light sources there were illuminating the image, what kind of light, and from what direction. The rules about what kinds of image manipulation are allowed often use some such phrases as allowing the lightening or darkening of pixels but not to the extent of adding or subtracting things visible or not visible. I argue that light sources outside the image are nevertheless things visible in the image. Including out of image light sources as things that must not be changed allows the lightening of a shadowed face, but not to the extent that suggests an extra beam of light falling on the face. I recall one controversially manipulated press image of a grieving funeral procession in a dim narrow street where many of the faces were so picked out as to suggest a gang of off camera assistants with torches picking out their faces. No solid thing was added or removed in that image, but the visual evidence of illumination sources certainly was.
Therefore I would like to propose that the selective combination of multiple exposures be permitted, provided they were all taken from the same position (say within a metre to allow for panorama swing) and the same time (say within a minute, to allow for changing filters or exposures). The intensity and colour of pixels can be changed considerably, but never in such a way as to introduce or remove things visible or not visible in the original scene, where "things" includes not only shadows or beams of light, but evidence of light sources from outside the image.
