RC2 technical questions

In a shameless attempt to move the discussion of RC2 into
a more technical arena, here are some interesting questions to
explore about RC2.
                --Bob

Key expansion
- How can you tell whether the permutation is based on
  some sequence of digits from PI?
- What are the diffusion and avalanche properties of
  this permutation?
- What are the linear characteristics of this permutation?
- What are the properties of the compression function
  that maps 16 bits (bytes X and Y) to 8 bits (byte Z)
  via Z = P[X + Y]?
- How does the length of the key influence the mixing
  of bits during each pass of the expansion algorithm?
  - Is this a non-linear feedback shift register over
    the field GF(256)?
- If the first pass of expansion is viewed as a hash
  function that produces 40 or 128 bits out, what are
  its properties?

Round Functions
- What are the diffusion and avalanche properties of
  the two round functions?
- What are the linear approximations and how good are they?
- What characteristics can be preserved by the round
  function that performs rotations?
  - With what probability?
  - Does the amount of rotation influence the security?
- What characteristics can be preserved by the round
  function that performs the data dependent selection
  of the expanded key?
  - With what probability?
- Are there any "weak" keys?  
  - Will the expansion algorithm produce them?

baldwin@rsa.com writes:
>         In a shameless attempt to move the discussion of RC2 into
> a more technical arena, here are some interesting questions to
> explore about RC2.
>                 --Bob
> 
> Key expansion
> - How can you tell whether the permutation is based on
>   some sequence of digits from PI?

[long list of other good and interesting questions deleted]

In a previous message, baldwin@rsa.com also wrote:
>WARNING NOTICE
...
>in such source code under applicable law, including without
>limitation trade secret and copyright protection.  In
>particular, RSA Data Security's RC2 (TM) symmetric block
>cipher source code has been illegally misappropriated and
>published.  Please be advised that these acts, as well as
>any retransmission or use of this source code, is a
>violation of trade secret, copyright and various other state
>and federal laws.  Any person or entity that acquires,
>discloses or uses this information without authorization or
>license to do so from RSA Data Security, Inc. is in
>violation of such laws and subject to applicable criminal
>and civil penalties, which may include monetary and punitive
>damages, payment of RSA's attorneys fees and other equitable
>relief.


Bob,

I'm confused by these two messages, as a non-lawyer (but I realize you're
also a non-lawyer).  How can RSADSI, on the one hand, expect to be able
to assert trade secret status over RC2 (with a warning to "...any person
who acquires, discloses or uses this information...") while at the same time
encouraging the world to examine and better understand the (illegally-
published) RC2 code?  To my lay mind, I cannot see how one can reconcile
your two messages.

I'm not trying to be cute or play lawyer.  I'm honestly confused as
to just what RSADSI's position here is.

-matt

> 
> I'm confused by these two messages, as a non-lawyer (but I realize you're
> also a non-lawyer).  How can RSADSI, on the one hand, expect to be able

	Giving Bob the benefit of the doubt here, I'm assuming that he
passed on the legal warning as a service to his employer, but he made
his post talking about RC2's technical strengths as an individual, not
speaking for his employer.

-- 
Sameer Parekh					Voice:   510-601-9777x3
Community ConneXion, Inc.			FAX:     510-601-9734
The Internet Privacy Provider			Dialin:  510-658-6376
http://www.c2.org/ (or login as "guest")		sameer@c2.org