It seems pretty clear to me. Both Jeff Schiller and Keith Moore have stated
<categorically that RC2, because of its "trade secret" status is not an
acceptible algorithm for an IETF standards track document. They get to
decide on that particular issue. I am fully in agreement with them on this
issue.
I don't know if RSADSI can still claim this to be a trade secret. Because of
the questionable status of RC4, the release of RC2 was handled very carefully
to make sure it was a genuine clean-room copy, and the whole process was very
publicly documented to ensure it could be verified by anyone. First an RC2
implementation was reverse-engineered, then a text-only specification of the
algorithm was written by someone on the other side of the planet, and finally
a third person (back on the other side of the planet again) who had never seen
the reverse-engineerd RC2 code wrote a new implementation based entirely on
the text-only specification. The entire process was documented in postings to
sci.crypt. Although IANAL, from what I've read of reverse-engineering this
should be good enough to withstand scrutiny - in any case a reasonably
sizeable non-US company which is using the code has had their lawyers look at
it briefly and is confident they'll win any legal challenge, at least based on
EU laws which AFAIK are mostly the same as US ones in this regard.
Peter.