[RC5] RC5 question

Zorba the Hutt zorbathut at uswest.net
Mon Oct 6 08:56:45 EDT 2003


How about NFSNET? http://www.nfsnet.org/

It's not doing anything *really* immediately useful, but it does have some
pure mathematical uses, and nobody ever knows what sort of things pure-math
ends up creating :P The only problems with it is that it uses a ton of RAM,
a ton of hard drive space, and the client is pretty bad. The first two
aren't really much of a problem, since it would be easy to (1) limit it to
high-speed computers, and (2) let people disable it if they want. The third
would be solved by using the distributed.net client :P

The request files are tiny - however, the result files are much larger,
which might cause some problems.

-Ben

----- Original Message ----- 
From: "Bruce Wilson" <bwilson at distributed.net>
To: "D.net Discussion" <rc5 at lists.distributed.net>
Sent: Sunday, October 05, 2003 6:03 AM
Subject: Re: [RC5] RC5 question


> Hi, Richard.
>
> I appreciate your comments, as you do raise excellent points.  I respond
in
> order to explain our position, and why things are as they are.
>
> >  BW> The point for distributed.net to participate in the contest is to
> >  BW> demonstrate the power of distributed computing
> > My Point is that one can demonstrate the power of distributed computing
> equally
> > well with projects that make much more sense to me than rc5-72+.
> >
> > For me making sense is to help mankind in some way.
> > (OGR helps to prove that the known shortest rulers ARE the shortest
> possible
> > climate prediction.net helps scientists to get insights into climate
> prediction
> > models
> > protein folding helps scientists to simulate how organic molecules fold.
> > ........)
>
> RC5 and OGR both have one very important advantage over the other projects
> you mention: they both have working cores written to work on our network.
> We have sounded the call in many forums (including this list) for coders
> with a passion for developing a new core (project of their choosing), and
> for projects who want to make use of our network, but we have not had many
> useful responses.
>
> We would love to have a whole family of projects which satisfy all the
> different definitions of "helping mankind".  We're looking and waiting for
> those projects.
>
>
> > I am not against the rc5 project.
> > I participated in rc5-64 and I run rc5-72 as a reserve project if the
> > servers run put of OGR stubs.
> > But I can't see any benefit in solving rc5 now.
> > What knowledge does mankind gain by solving rc5-72 that it doesn't have
> now ?
>
> As our primary goal is still distributed computing, we do gain knowledge
> about distributed computing.  Yes there are other paths to this, but the
RSA
> contests serve as a good "standby project" which can keep our network busy
> as other short-term projects come and go.
>
> And as I mentioned earlier, the RSA contests have something no other
project
> you've mentioned has:  a cash prize for completion.  We have a total
annual
> operating budget of $0, aside from donations and contest winnings.
> Completing the contest and getting to keep $2000 or $6000 is very
> attractive.  Likewise, the chance of winning $1000 or $2000 helps to draw
in
> new participants.  If those participants choose to back-burner the RC5
> projects, we gain more support for whatever other projects we are running.
>
>
> >  BW> Participating in a crypto contest happens to be a good way to apply
> >  BW> distributed computing technology.
> > Yes.
> > But I am personally sure that there are projects that a) are well
> distributable
> > and b) make some sense by providing new information to all of us.
>
> But until c) someone who understands them joins us to help author a core,
we
> can only talk.
>
>
> >  BW> Various people within DCTI have various opinions about the
usefulness
> >  BW> of the results of the project, politically and otherwise, but
> >  BW> ultimately we're not making any statement other than "distributed
> >  BW> computing can do that".
> > The great thing is that distributed.net offers a client where you can
> simply
> > plug in new projects.
>
> There is a little more to it than simply plugging it in.  We also have to
> make the projects workunits fit into our distribution format, accomodate
key
> distribution at the master and proxies, build in some anti-cheating tests,
> and come up with a meaningful way to display stats.  For a project which,
> for example, requires 256K of data transferred per work unit, we would
need
> to make some significant changes to the entire chain.
>
> Still, all these issues can be overcome.  We just need a reason to make
the
> change.
>
>
> > The reason for my posts is to get people started to think which of the 2
> > current projects they set the focus on.
>
> And for some people, RC5-72 will still be the right choice.  Not because
of
> any particular message they want to send, but because they know that every
> block they process has an equal likelihood of containing the winning key,
> earning them the prize.  Others like yourself prefer OGR because it has
> tangible benefits to mankind.  Some only do it for the stats, so they will
> choose the project to which their platform is most suited.  Others just
want
> bragging rights.  We welcome all these participants equally.
>
>
> > Maybe there will be a third project somewhen in the future.
> > (maybe dnet could cooperate with some biochemistry institute ...
> > just dreaming :)
>
> The invitation is open.  If you can point such a project to us, or point
us
> to them, we would be glad to talk about ways we can work together.
>
> _______________________________________________
> rc5 mailing list
> rc5 at lists.distributed.net
> http://lists.distributed.net/mailman/listinfo/rc5
>



More information about the rc5 mailing list