Some initial comments on the draft, mostly nits.
- The expanding integer formats used by new-style CTBs and by version-4
signature subpacket fields should probably be given names and added to
the catalog of types in section 3. The two formats are slightly
different; what is the reasoning behind this?
- Section 2.4.1 does not fully describe the CRC used with ascii armor.
- Section 7 does not mention the possibility of a subkey packet occurring
in a transferable public key... I assume this is just an oversight.
- The draft does not describe how session keys and hash digests are
encoded into MPIs for use in public-key algorithms. IMHO this is the
most glaring omission.
Wim Lewis / wiml(_at_)hhhh(_dot_)org