My understanding is the opposite. RFC 2822 seems to me to imply that
the Reply-To address is a substitute for the From address - i.e.(b).
I would argue that RFC 2822 propagated the error in RFC 822. But it's
pretty much irrelevant now, because (b) is the widespread behavior.
No, RFC 822 is/was a Standard. Standards do not, by definition, contain
"errors". "Bugs" maybe, but not errors.
Okay, this is getting very semantically picky, but to me:
A mistake in the specification is indicative of an error of some kind on
the part of the specification authors or editors.
A bug is a failure to conform to a specification.
But if we want to argue about the meanings of words and who shall be
master, we should just quote Humpty Dumpty (via Lewis Carroll) and be
done with it.
If that were not so, then anyone would be able to cherry pick which bits
of a standard they were free to ignore.
In fact, anyone *is* free to cherry pick which bits of a standard they wish
to ignore. There are no standards police, no prisons or death penalty for
standards violations. (though occasionally I wish there were ...)