At 08:58 PM 7/21/99 +0100, Brian J. Beesley wrote:
>On 20 Jul 99, at 0:26, Luke Welsh wrote:
>
>> Eventually, PrimeNet should be able to move small (and large)
>> amounts of money around the world with ease.
>>
>Unfortunately, e-cash opens several other cans of worms!
[....]
>what happens if the winner comes from an "unfriendly" country?
>Can we transfer e-cash to a citizen of, e.g., Iraq,
How does this differ from the current situation? What if
Nayan was North Korean? The can of worms is already open.
GIMPS/PrimeNet would not be the only organization in the
e-cash boat. It is a problem to be worked out by the big
boys. (Anyway, if we're in that boat, at least we have our
can of worms to fish with :-)
>> We really do need a formula for [e]. Set aside $25,000 and each
>> x% improvement in GIMPS throughput wins x% of the $25,000?
>
>What's wrong with having a panel (possibly consisting of previous
>Mersenne prime discoverers) to evaluate any contenders for this &
>judge how much, if any, of the fund should be awarded for each
>improvement?
Sensible idea. The panel could come up with their own set of rules.
I'm not sure I'd want to be on such a panel unless my vote was
anonymous ("Sorry, Heloisa, but I think your idea was worth only....").
>> Award $10,ooo for each new Mp, regardless of size. Suppose M(7777777)
>> is prime. The finder is credited $10,ooo (to be awarded when GIMPS
>> finds a decamega prime). The decamega finder also gets $10,ooo.
[...]
>This would seem to be a bit unfair. I think the decamega finder
>should definitely get a large share (25% minimum).
See remarks, below.
>The point here is
>that, if I can freely download a program & use it to check Mersenne
>numbers in the 10 million digit range, why should I bother to
>contribute to a cooperative project if, by doing so, I'm going to
>have to give up almost all the prize if I get lucky?
We didn't have this problem when there was no prize. Can't we simply
promote it as a $10,000 prize, not a $100,000 prize?
My counter-argument is Distributed.net. They are going for the
US$10,000 offered by RSA Labs.
http://www.distributed.net/rc5
10% of that will go to the winner. And I *think* they are bigger
than GIMPS. S@H is much bigger.
>The fact that
>even poor people are prepared to throw a few dollars at the remote
>chance of a lottery jackpot would tend to indicate that a few dollars
>recompense for CPU cycles contributed to a cooperative project will
>not attract contributors in the way that a large cash prize will.
Do we have a feel for how many contributors have been attracted solely
by the large cash award ?
>I'd prefer to keep a slice (smaller than the discoverer's slice) for
>discoverers of non-qualifying Mersenne primes, but don't pay anything
>out until the range up to 10 million digits is (more or less)
>completely searched. Then divide it equally between them. Carry the
>sum forward if it happens that we discover that we already know all
>the Mersenne primes with less than 10 million digits.
It might be years before the range has been completely checked.
Double checked? What's the phrase? "a sticky wicket"?
>> Is $10,ooo about right?
[....]
>.. if we try to link the amount of the prize to income, then how
>much would we have to pay Bill Gates if he won it?
Exactly my point..... "Is $10,000 about right?"
I still favor something along the lines of:
[a] $10,ooo for decamega Mp finder
[b] $20,ooo to be split amongst all other Mp finders
[c] $25,ooo for algorithm improvements
[d] $45,ooo for 1st LL, 2nd LL, factors, Cunningham factors
There's another reason I like [d]. If [d] were in place, I'd
stand a better chance of convincing a couple large companies to
come onboard GIMPS and funnel their earnings to charity.
Now Gordon and Glenn see things differently.
At 04:50 PM 7/22/99 +0100, Gordon Spence wrote:
>In fact the answer is simple, the person who discovers it is
>*entitled* to it, end of story.
>I would hope that they would *choose* to donate some to George
>and Scott, but if they decide not to well then that is up to them.
I would also hope so, but I think this is unlikely... taxes, family
pressures, etc.
>Direct quote from eff page
>> Through the EFF Cooperative Computing Awards, EFF will confer prizes of:
>> $50,000 to the first individual or group who discovers
>> a prime number with at least 1,000,000 decimal digits
>
>Seems clear to me, the GIMPS s/w is given away free to the individual
>who runs the test.
"individual or group". It seems clear to me that the award should
go to the group, since this is not an individual effort. I think
eff.org was very deliberate in their language.
>Now that would be interesting, what would the EFF do then given that
>they are entirely about freedom and the rights of the infividual.
I think they foresaw many of the amicable disagreements we are experiencing.
I just *knew* this award business would be a "tar baby" :-(
>As a Mersenne prime discoverer it would appear that I would be on the
panel(s)
>that have been suggested, and my vote goes to allowing the discoverer to keep
>it and do what their conscience tells them is the right thing to do.
Do you get one vote, or 1/2? Do I split my vote with Walt Colquitt?
Can Walt delegate his half to me? Does Alex Hurwitz get two? How many
votes does George Woltman get? Landon Noll is disqualified. Does
Arial Glenn get his vote(s)? Vote by majority or plurality? What
a mess.
>I don't agree with Luke's sentiments about orderly progress, having
>already tested one number in the 20m range.
OK, let's talk about this.
>As long as they all get tested does it matter in what order?
But they all won't get tested, not for another 20+ years. After the
10Mdigit prime, there will be the 100Mdigit prime, then the giga-digit
prime. GIMPS has been lauded for conducting an orderly search, even
more so for double checking. Jumping ahead, the search space will once
again become horribly fragmented -- a giant step backwards into the Cray
era.
Having said that, I am confident that there will be nothing in V19 to
prevent one from jumping ahead. I just don't think it should be
encouraged. And I agree, it is a fine line to draw.
To quote eff.org again:
"EFF hopes to spur the technology of cooperative networking..."
Perhaps another collective will form, an modern version of the Amdahl
group .... just look for the World's Largest Prime, one that has
10 million digits? Since we know there are an inifitude of primes, I
do not find such a search very interesting. I would wish them luck,
but it's just not something for me.
At 02:55 PM 7/22/99 -0400, Glenn McLaren wrote:
>I agree with Gordon.[...]
>The moment you start splitting the prize is the moment someone else
>will release software in competion to us,
I think we will see another group form regardless. I will not join
them as I do not think their discoveries will contribute very much.
Others will disagree and that's OK.
So I respectfully stand by my position.
>Gordon Spence (finder of M2976221 which earned me precisely $0.00 <G> )
I, however, garnered a free pitcher of beer! Unfortunately, it was
mass produced American beer, so I think you got the better deal.
--Luke, who'd sure like a Schneider Weiss right now :-(
_________________________________________________________________
Unsubscribe & list info -- http://www.scruz.net/~luke/signup.htm
Mersenne Prime FAQ -- http://www.tasam.com/~lrwiman/FAQ-mers
