PRZ style CFB mode description preferences

[Adam Back <aba@xxxxxxxxxxxx>]

Some time ago Hal asked which of two descriptions for the CFB mode
Phil Zimmermann invented were preferred.  (The CFB mode in summary
does not use a fixed feedback length in bits, as is normal, but varies
that feedback length at semantic boundaries in the plaintext).

My preference is greatly for the description referring to the bulk of
the data encryption as CFB-64, and the re-syncs as having CFB-16
applied to them.  (Or CFB-whatever).

Adam