Just to clarify...
Speaking personally, while I support the idea of non-disclosure of the
public key, if I thought it represented unnecessary complexity I would
vote against it. Let me observe that the ability to do this "falls out"
of the chosen identifier mechanism, in particular arbitrary key
selectors.
I said this is a circular argument since the arbitrary key selectors
wer chosen in order to hide the public key. Strictly speaking, the
key selector may originally have been chosen (for all I know) to solve
the problem that everyone might not use MD5 for a public key digest.
But also on the table at the time was the realization that we don't
need to save bandwidth with a digest, and using the public key iteslf
as the identifier would be okay. In my mind, the arbitrary key
selector stands in opposition to this, and so my point that the
argument is circular still stands.
- Jeff