Jon Callas <jon(_at_)pgp(_dot_)com> writes:
The "hand huffman-coded" structures go all the way back to the origins of
PGP, they're hardly new.
There are a number of bit packing methods which were not in
pgp2.x.which are in the OpenPGP draft:
- 8 bit floatpoint format for S2k iteration count
- new 192 format
Someone suggested that these were mostly decided prior to company
formation -- which suggests that Colin Plumb, Derek Atkins are the
guilty parties (and perhaps PRZ).
I agree with the sentiment that it would be nice to have a system
laid out with long constants, but I think that that would be a good
thing to do in V2 of the spec. I'd like to get this one done, and
then let the working group debate new formats.
This (saving 32 bit constants for OpenPGPv2) only makes sense, and
will only really be possible if the plan is to have an OpenPGPv2 which
only has SHOULD for backwards compatibility options. Is this the
plan? That OpenPGPv2 should be the clean up version?