[RC5] RC5 question

Décio Luiz Gazzoni Filho decio at revistapcs.com.br
Sat Oct 4 00:07:59 EDT 2003

Hash: SHA1

On Friday 03 October 2003 20:10, Daniel Quintiliani wrote:
> You just don't look at it as a contest. There were no prizes for SETI, or
> for our OGR, and still they're fun. I ran SETI for 3 years up until last
> summer. I knew the likelihood of me finding ET, and when the project
> finished I noticed that my name wasn't on the list of good workunits, but
> it was still a fun thing to do. (I did get annoyed later when I heard that
> they were going to delete the statistics.) Anyway, these things are just
> fun to do. I'm just suggesting that 128-bit would be more fun than 72-bit
> since it would be more relevant today.

You're looking at it from the wrong perspective IMO. For instance, the  
finding of ET would be the most important discovery of mankind in a long 
while, if not ever. To some, it justifies all the computing power that you 
throw at it, even if we never manage to finish it. By the way, I dispute your 
claims about knowing the likelihood of finding ET -- nobody does. Now there's 
a clear analysis that can be applied to an RC5 key search, and all odds are 
against it. As I argued in another post, by the time we find it out (if 
ever), it won't be relevant anyway.

I recognize two schools of thought in the DC community: 1) the project may not 
be all that important, but it can be done, 2) the project is extremely 
important, and even if the odds are against us, it's worth the try. RC5-72 is 
an example of the former, SETI is an example of the latter. Maybe there are 
some important _and_ doable projects, like disease research, but the truth is 
that math-related projects (crypto, GIMPS, SoB, ECMNET, NFSNET) have been the 
only success stories of distributed computing. Now you're proposing a 
non-important project which can't be done -- for all we know it's less likely 
than finding ET -- and I just don't see the point. But as I mentioned in the 
other post, there were people rallying behind The Neo Project's beyond-futile 
attempt at factoring large integers by brute force, so it wouldn't surprise 
me to find out that there people on the DC community that support that idea, 
though the rationale behind it is, and will always be, beyond my grasp.

I see it this way: cracking a 72-bit key, even if not exactly relevant (I'll 
not go into it, but it's more relevant than you think it is), at least it's 
_doable_, which I believe is far more important than the supposed relevance 
of a cipher that we'll never break anyway -- that being RC5-128.

By the way, why don't we run OGR-150? It would probably be more relevant to 
the real-world applications of OGR, no?

Version: GnuPG v1.2.3 (GNU/Linux)


More information about the rc5 mailing list