Frequently asked questions in sci.crypt.research answered.
Last-modified: 14 August 2002
-----BEGIN PGP SIGNED MESSAGE-----
Updated 14 August 2002
Last change: updated export regulation information
1. "What is the charter of sci.crypt.research?"
The discussion of cryptography, cryptanalysis, and related issues in a
more civilized environment than sci.crypt. In particular, we are more
interested in the technical aspects of serious cryptology. This is a
moderated news group. Before posting, you may want to consider if your
post would be more appropriate in talk.politics.crypto (discussions of
the relationship between cryptography and government), sci.crypt
(technical discussions of cryptography, unmoderated), alt.security.pgp
(discussion of Philip Zimmerman's Pretty Good Privacy program and
related tools, programs, and issues), alt.security.ripem (Mark Riordan's
Privacy Enhanced Mail program), alt.security (general computer security
issues), or some other group.
2. "How do I submit an article to sci.crypt.research?"
Most news posting software will recognize sci.crypt.research as a
moderated news group and redirect submissions to the submissions address.
As an alternative, you can send your article directly to
email@example.com for consideration.
3. "What do you think of my new cryptosystem?"
GUIDELINES FOR POSTING NEW ENCRYPTION SCHEMES TO SCI.CRYPT.RESEARCH
People frequently invent new encryption schemes and protocols and want
to share the fruit of their creativity with other people sharing an
interest in cryptography. Past experience on sci.crypt indicates that
many of these postings tend to be just an annoyance, rather than serious
research. In an attempt to cut down on the annoyances, while still
encouraging serious research in this area, we have proposed the
following guidelines for posting new algorithms.
A. DO research other encryption methods and understand how they work,
including both historical and current work. There are lots of good
books and journals devoted to this kind of work.
B. DO investigate methods of breaking encryption algorithms, or
cryptanalysis. Knowing how a cryptanalyst might go about trying to
break a cipher gives you much better insight into how to create a good
one. Indeed, among professionals, experience attempting to break
encryption methods is considered essential before designing new ciphers.
C. DO COMPLETELY DOCUMENT your algorithm with both a text description
and, if applicable, computer source code. By "completely document" we
mean that the description is sufficient for anyone skilled in the art to
implement or simulate your algorithm. If you have doubts about export
restrictions on the source code for the algorithm, you may choose to
provide a pointer to a place where the source code can be obtained by
qualified people, rather than posting it. If you have a complete
application using encryption, and you are posting from the USA or
Canada, then providing a pointer to the program rather than just posting
it is recommended, but the text description should still be posted.
D. DO describe the advantages of your algorithm compared to others in
existence, including comparison of efficiency and other relevant design
parameters. Make sure that you provide evidence to support your claims.
E. DO try to break your own scheme before you post it. This could save
F. DO take a look at similar postings from other people on sci.crypt
and sci.crypt.research and try to analyze them. This will give you some
insight into how others will look at your posting and perhaps allow you
to make yours more clear. It also gives you a chance to try to break
some other cryptosystems.
G. DO read the sci.crypt FAQ, posted monthly and archived at
rtfm.mit.edu under /pub/usenet/sci.crypt.
H. DO describe which quantities in your scheme are public and which are
private. Explicitly mention what the key is and what the message is.
I. DO include the design principles you used and mention any assumptions
you made which you think may be relevant. Explain why you think your
system is secure.
J. DON'T expect a response from a ciphertext only ("Try and break
this") challenge. Although there are techniques for attacking
ciphertext only, most of them require lots of examples, some of which
correspond to known plain text. They are also rather time consuming.
If you do feel the urge to issue a challenge, you should make sure your
posting complies with all of the above guidelines. Offering a cash
reward if someone breaks your cryptosystem may help someone to be more
motivated to try (and is also a good test of how much you believe in
your own system).
K. DO include the information covered in the points above in your
posting, or at least include enough to allow people to evaluate your
scheme. DON'T post a message containing ONLY a URL and little more.
L. Be ready to carefully evaluate and learn from any feedback you get.
4. "What effect do export regulations have on this group?"
You are advised to familiarize yourself with the current export
regulations pertaining to your country. In the USA, a good starting
place is at http://www.bxa.doc.gov/Encryption/Default.htm. Most postings
to this group are international academic discussions pertaining to
cryptography and cryptanalysis that are protected as free speech and
free publication by the U. S. Constitution (in the USA), and are not
restricted from export. In the USA, cryptographic source code relavant
to a discussion in this news group may be required to be reported as
Discussions of export controls are considered "off topic" for this
group, and are better posted in talk.politics.crypto.
Comments, questions, or suggested additions to this FAQ should be
directed to the sci.crypt.research moderators at
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.0.7 (Cygwin32)
-----END PGP SIGNATURE-----