The cornermask is, a you mentioned, an offense against the PI philosophy.
It depends on how and why you apply it. If by "PI philosophy" you mean my philosophy, then here it is. If you:
- Analyze the data to find and outline the problem,
- Understand the problem and its causes (for example, an ampglow artifact or a light leakage issue in this case),
- Design a suitable solution to the problem (a local mask in this case),
- Evaluate the side effects and the risks of your solution (decreased SNR on the masked area in this case, and the risk of introducing a nonuniform illumination artifact),
- Make an informed decision as to whether the solution of the problem is worth the side effects and risks,
- Apply the solution strictly to solve the problem, preferably in a purely algorithmic way,
- Evaluate the result of the applied solution and its side effects,
then what you are doing is astrophotography: Turning data into information to communicate something from nature. However, if you do one or more of the following:
- Apply arbitrary manual retouching techniques without documentary criteria (example: create a layer mask, paint it with a brush by hand, then sharpen where you want because it looks nicer that way),
- Remove objects or image structures selectively, without documentary criteria (for example, removing all the stars in a wide-field image can be a nice way of revealing the true shape of a nebula, but removing a large blooming by cloning its adjacent pixels is like "inventing" that part of the image),
- Add artificial elements or artifacts "for aesthetic purposes" (example: artificial star spikes),
- Apply any process selectively without documentary criteria, i.e. without criteria based on a thorough analysis and understanding of the properties and nature of the data and of the objects represented (typically also for aesthetic reasons),
- Apply processes incorrectly, or without the necessary knowledge and common sense, (typically, trying to process an image beyond the reasonable limits imposed by the SNR of the data),
then what you are doing is, to a lesser or greater extent, painting. Downloading the image from the Internet to modify it, or painting it directly on a canvas, may be easier than acquiring it from the sky, and the results may be aesthetically better, too.
Astrophotography, as I understand it, is much more about the "how" and the "why" than about the results. When I look at an image, I need to be sure that I can trust what I am seeing. For this to happen, I need to know how the image has been processed and why. In essence, this is also the philosophy behind PixInsight, but everybody can use a tool as they want.