[rc5] Official Statement from distributed.net
nugget at slacker.com
Tue Jun 24 01:27:55 EDT 1997
Everyone here at distributed.net has been so focused on completing the V2
clients and keyservers that I realize we've let several issues go
unresolved for some time now. Now that the coding has reached the final
stages, we've been able to take a short breath and refocus on the entire
project and the direction we see it heading.
With the sudden fall of DES and the steady influx of fresh new faces, we've
been doing all we can to simply react to the increased demand for the
clients, information, and statistics. If we have in any way failed to
answer YOUR questions in a timely fashion, I'd like to apologize here and now.
There has been quite a bit of traffic here on the list lately, and I'd like
to take this opportunity to address each question with the best answers I
can provide at this time. Hopefully this will answer the majority of your
questions and put most (if not all) of your fears to rest.
1. Migration from the Version 1 clients to the Version 2 clients
It has always been our goal at distributed.net to focus on and beyond the
56-bit RC5 contest in general. To that end, we are trying to build a
less targeted code base for our client software. Rather than build the
best RC5 mousetrap imagineable, we are investing the effort in building
a more general-purpose distributed computing engine that can be rapidly
and efficiently modified to support any number of distributed computing
tasks. Many of these have been mentioned here on the list, such as the
GIMPS Merseinne Primes Search, and the SETI data analysis. It is our
goal to build distributed.net into a network of participating peer clients
used to tackle any number of these daunting tasks, and the building of
the Version 2 (and eventually Version 3) clients is the first step towards
WE HAVE NO INTENTIONS OF REMOVING SUPPORT FOR THE VERSION 1 CLIENTS.
The clients and code you are running today will continue to run and there
are no plans to change this.
It is apparently misunderstood by many that our primary goal in building
the V2 clients is to "tamper-proof" the code and protocol to prevent
sabotage of the effort. This is not the case. Anyone who has spent
even a small amount of time looking at the old code will realize that
although it is functional, it is certainly not as scaleable and upgradable
as it should be. In order to move this effort the direction we envision,
a fresh code base was required; hence V2.
We do not presume that simply by obscuring the code we will make the
effort safe from intervention. Rather, those protections that are in
place now, and those that are planned, will be server-side and will
be in effect for both V1 and V2 clients alike.
2. Availability of the V2 source code
We do retain the right to significantly limit the release of the source
for the V2 clients. If your site requires the availability of the source
in order to participate, the V1 clients will be available and you will
be able to maintain your support of the effort exactly as you are today.
Our reasoning behind this decision is that the V2 clients are in reality,
just an intermediate stage between here and our final goal of a truly
flexible remote computing client.
Much of the confusion and difficulty we are experiencing right now is a
direct result of the parallel client development that has taken place
regarding the V1 clients. It is our desire to centralize all coding and
development effort from this point on. The only way we can guarantee
this is to limit the distribution of the development.
It should be mentioned that anyone is welcome to join the coding effort.
We recognize the considerable talent that was present in the DES efforts
at large and certainly understand how much we all can stand to benefit
from the experience that those efforts have gained. There are some
outstanding coders out there who have not yet had the opportunity to
contribute to this effort.
3. What is taking so long on the v2 clients?
The V2 clients are effectively finished at this point. We are, however,
having to upgrade the keyserver network to handle the dual-protocol
traffic, and obviously cannot release the clients until this has been
completed and tested to everyone's satisfaction. The extra time spent
in testing tonight will simply prevent the hours of headaches were we
to release the clients prematurely.
Again, I apologize for any apparent lack of availability amongst the
distributed.net organizers. We have been so focused on some very specific
issues this past week that it has been very easy to lose sight of the
entire effort. Also, we were not prepared for the incredible swarm of
successful DES crackers. Those efforts simply were not supposed to finish
If you have any questions, please DO NOT HESITATE to contact myself, or
any of the other distributed.net organizers. We are always open to
suggestions, donations, and volunteers. It is only through efficient
cooperation that we will be able to pull this off successfully.
nugget at slacker.com
To unsubscribe, send email to majordomo at llamas.net with 'unsubscribe rc5' in the body.
More information about the rc5