The sense of Small Caps could be to have the lower case in the form of upper case and the sense of Petite Capitals to have the upper case in the size of the lower case (more or less).
As long as it is |A| and |a|, the situation is clear.
But when am I dealing with glyphs that seem to have no such alternative? I refer for example to U+01BB, U+01BE, U+01C0-3: how are these glyphs to be understood? Does it make sense to produce a .sc or .pc version?
I see that in Brill font there are many glyphs rendered as SmallCaps, but they don't run out of all lowercase letters. On the basis of which criterion was the choice made?
In the case of the Brill types, I provided smallcap variants for a lot of lowercase only characters because of the nature of Brill’s publishing, especially in the area of linguistics, and their typographic use of smallcaps in running headers and subheads. My general approach was to provide smcp lowercase-to-smallcap mappings for any characters that occur in natural orthographies, as distinct from those that occur only in phonetic transcription notations (and yes, there may be some inconsistencies where I was doubtful about the use of a character and erred to include in smcp).
To make things easier, I copied the block below as comma-separated names. Please note that .da (design alternate) and .la (language alternate) are special variants I use, but are probably unnecessary to anyone else.
The IPA Extensions block:
And what we actually have here:
Similar verification needs to be made in all Latin and IPA blocks.
Finally, a word about rotated or mirrored glyphs: you can use a component flipped or rotated if you have a stress of 0° (like the captures above) or the font is a sans with no contrast. But this will fail for other design styles because the contrast will be reversed. To get perfect glyphs, there is no scape from laborious work.