Eliminate experimental modes

Every time when I open Pillow's C sources, I absolutely scared by all these experimental modes. Do I need to worry how not to break them? Do I need to implement features for all of them? How can I write tests for them?

In my opinion, an experiment must have the following characteristics: hypothesis; subject; deadline or exit condition; metric; outcomes. When someone adds another else if in ImagingNewPrologueSubtype, this is not an experiment. This is just another unfinished feature which someone else needs to support.

My suggestions are:

• Remove following modes with IMAGING_TYPE_SPECIAL type: I;16, I;16L, I;16B, I;16N, BGR;15, BGR;16, BGR;24, BGR;32
• Remove RGBX mode as full equivalent of RGB
• Keep widespread 1, P, L, LA, RGB and RGBA modes
• Keep alternative linear (or near linear) color spaces CMYK and YCbCr
• Keep derivatives with multiplicated alpha La and RGBa
• Apparently, keep wide modes F and I
• Create full list of the supported modes, monitor correctness of new or modified operations with all of them and include tests

Questionable modes:

• Alternative color spaces with signed channels LAB and HSV
• Palette with alpha PA. As I know palette supports RGBA modes, so it can handle "palette with alpha" in a different way

Start by reviewing Pillow's C sources and the ImagingNewPrologueSubtype entry point, then inventory the listed image modes and their existing operations. Determine the supported-mode list and test coverage needed before removing or retaining modes; done means the mode policy is agreed and correctness is covered across all retained modes.