> Is the idea that there's some inherent mistrust of `-c copy` or that sometimes downstream options affect it basically invalidating it?
Yes, that's one reason.
I suppose the main mistrust along those lines is -- I have all these programs which manipulate my media metadata and sometimes changes the names or locations of my media files. And I'm basically fine with lots of small automated changes to my metadata from programs like `beets`. I'd just like some assurance whatever they spit out is what I started with.
With respect to metadata more specifically, if someone cleans up the metadata on an album or adds additional information, or album art, this shouldn't invalidate any checksum.
Network transfers of media could certainly benefit from this. If I send a ALAC album to someone, and they open it 3 months later, they should be able to know what I sent is what they are listening to, even after they retagged it.
