A question about Camellia

[Doug Barton]

Robert J. Hansen wrote:
> Faramir wrote:
>>   Well, I don't think you are crazy, but I am part of the group that
>> likes to be able to chose between several options, provided all the
>> options are secure.
> 
> That "provided" is the sticking point.  Small is beautiful, IMO.  YMMV.

I agree that small is beautiful. I also think that there is something
to be said for pruning seldom-used code paths in a security-related
application since there is always the possibility of bit rot, as well
as longstanding bugs that have not yet been identified yet because the
code has not been exposed to a wide enough variety of use cases.

OTOH, in something like PGP there is a definite need to maintain
backwards compatibility. There is also the issue of needing options
that can easily be exercised with existing tools should a current
"favorite" be found to be insecure down the road.

My personal opinion is that gnupg treads this line pretty well. If it
were up to me I would like to see more configure --options to not
include some of the options I'm not likely to ever need, but I don't
care enough about it to put the work into making a patch. :)

> The Air Force was shocked by this and canceled their cooperation in the
> program.  They learned from the F-4 and the F-15 that twin-engine
> aircraft had more than twice the engine problems of single-engine
> aircraft.  The downside of the extra complexity was greater than the
> upside of having a second engine. 

Regardless of whether this story is apocryphal or not, there is
something to be learned here. The AF's conclusion seems obvious,
however it ignores a critical factor of the Navy's use case. Because
most of the time Navy jets operate over water (I would hazard a guess
to say that it's a significant majority of the time, but I'm not an
expert), and because having to eject over water has a much higher
fatality rate than ejecting over land, "getting back to the carrier"
is much, much more important for the Navy and (likely, again, not an
expert) justifies the additional complexity. The lesson here being,
make sure you understand ALL the parameters before you make your
conclusions.

BTW, to get back to Camellia, I had to do some research on this in
another context and while I'm not prepared to judge the "safety" of
the cipher, I did come across this post from David which I thought was
important. Short version, don't use this for production stuff since
things you encrypt now might be undecryptable in future versions of
gnupg.

http://www.nabble.com/Camellia-td11158599.html


hope this helps,

Doug